Google Authenticator vs 2FAS

Agent access

1. ai-native userPoint an agent at llms.txt or agent-oriented docs

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS✓fullprobed9/10

   Direct probe confirms a live llms.txt with an explicit 'For AI agents' section, plus modular per-section llms files, llms-full.txt, per-page .md mirrors, and Accept: text/markdown rendering — strong first-party agent-oriented docs infrastructure verified via runtime probes. Missing for 10: no independent/community corroboration of agents actually consuming these docs in practice.

   • [probe] “PROBE llms.txt: HTTP 200 at https://2fas.com/llms.txt # 2FAS — Password Manager and Authenticator > Local-first password manager (2FAS Pass…”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE runtime (recorded 2026-09-15): machine-readable pricing verified live at https://2fas.com/.well-known/pricing.md — '# 2FAS Pricing... …”

   proof ↗⚑ flag

2. ai-native userRun the product headlessly / in CI for automation

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a mobile authenticator/password manager with no CLI, headless mode, or API for CI/automation; the product's own llms.txt explicitly declares 'Public API: None' and is local-first by design, meaning it cannot be run headlessly or scripted in CI.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

3. ai-native userConnect an agent via an official MCP server

   weight 3 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a password/2FA manager, not an agent, so an MCP server axis could plausibly apply to its ecosystem; however evidence explicitly states 'Public API: None' and no OpenAPI/MCP endpoint exists, with probes confirming no API surface for agents to connect to.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

4. ai-native userUse an official CLI

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a mobile/browser 2FA and password-manager app with no evidence of an official CLI; docs explicitly note 'Public API: None' and no CLI tool is mentioned anywhere in the evidence pack.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

5. ai-native userDrive the product through a documented public API

   weight 3 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS explicitly states 'Public API: None' in its own llms.txt, and OpenAPI/swagger probes all return 404, confirming there is no documented public API to drive the product programmatically — it is local-first by design.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

6. ai-native userIssue scoped/least-privilege API credentials for an agent

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a local-first authenticator/password manager with no public API; its own llms.txt explicitly states 'Public API: None' and no agent-callable endpoint for token generation, so there is no mechanism to issue scoped/least-privilege API credentials for an agent.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

7. ai-native userBuild against official SDKs

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS explicitly declares 'Public API: None' and has no OpenAPI/SDK endpoints; probes confirm no SDK or API surface exists for developers to build against.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

Agentic features

1. ai-native userSet up automations that run autonomously in the background

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

2. ai-native userOperate the product with natural-language commands

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

Api quality

1. ai-native userRely on versioned APIs with a documented deprecation policy

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS explicitly declares 'Public API: None' and has no OpenAPI/swagger endpoints (all probes 404), meaning there is no versioned API at all, let alone a documented deprecation policy.

   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

1. ai-native userPerform bulk operations across many items at once

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a local-first authenticator/password manager with no public API (explicitly declared 'Public API: None'), and no evidence of bulk operations across items (e.g., batch import/export beyond single-file import, bulk edit/delete of tokens or vault entries) that an AI agent could invoke. There's no scripting, CLI, or API surface for programmatic bulk actions.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [claimed-docs] “We support imports from Aegis, Raivo, LastPassAegis, Stratum (formerly Authenticator Pro) and Google Authenticator.”

   proof ↗⚑ flag

1. ai-native userDo everything through the API that I can do in the UI

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS's own llms.txt explicitly declares 'Public API: None' and probes confirm no OpenAPI/swagger spec exists at any candidate path, meaning there is no API surface at all through which a user could replicate UI actions.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

2. ai-native userExport all of my data in open formats and leave

   weight 3 · round to 2FAS
   Google Authenticator!disputedcontradicted3/10

   Docs show a QR-code based transfer/export feature for moving codes between devices (docs-3, docs-11, comm-9, comm-18), but this is a proprietary re-import-only format tied to Google Authenticator itself, not an open, portable format usable elsewhere. Community evidence explicitly contradicts any 'open format export' claim: users report Google 'makes it stupendously difficult to just get a simple text export to a file' and that 'Google Authenticator does not allow you to include your codes in your backup, even when it's encrypted,' with data tied to a Google Account that can vanish if banned. Missing for 10: any documented open/standard export format (e.g., plain TOTP URI list, CSV, encrypted portable file), and evidence of true account-independent data portability.

   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “To manually transfer Authenticator codes to a new device, you need: Your old device with Google Authenticator codes”
   • [community] “Storing it in the google cloud doesn't satisfy me. I just simply want the codes under my control... google still makes it stupendously diffi…”
   • [community] “Google Authenticator does not allow you to include your codes in your backup, even when it's encrypted. So due to Google Authenticator, losi…”
   • [community] “Google Authenticator already has a QR-Code based very easy export procedure, I just backup my GAuth to my spare phone and tablet. It feels s…”
   • [community] “Google Authenticator has a feature to allow export of all saved codes into one QR code for re-import into Google Authenticator running on an…”
   • [community] “If Google ever decides to kick you out of your account, Authenticator data will be gone. Google has done this on several occasions in the pa…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS~partialcommunity6/10

   2FAS explicitly supports exporting/backing up data (encrypted export files, self-host freedom, import/export in free tier) and even imports from competitor formats, showing some portability. However, there's no evidence the export format itself is documented as an open/standard schema (e.g., JSON spec or standard otpauth migration format), and the backup is typically encrypted for 2FAS's own re-import rather than a documented open format for third-party tools. missing for 10: explicit open/standard export format spec, independent confirmation the exported file is usable outside 2FAS ecosystem, and details on the exact format encoding.

   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [claimed-docs] “Freedom to self-host your data”
   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [community] “Free tier of 2FAS Pass includes auto-fill, import/export, and standard password generator; paid tier adds multi-device sync, more than 200 v…”
   • [claimed-docs] “We support imports from Aegis, Raivo, LastPassAegis, Stratum (formerly Authenticator Pro) and Google Authenticator.”

   proof ↗⚑ flag

3. ai-native userRead the product's source under an open license

   weight 2 · round to 2FAS
   Google Authenticator~partialclaimed4/10

   The official GitHub repo (google/google-authenticator) publicly hosts source for HOTP/TOTP implementations across platforms and a PAM module, suggesting some source is open, but the evidence never states an explicit open-source license nor confirms the current mobile app's source is included/maintained there. missing for 10: explicit license text (e.g., Apache 2.0) confirmation, evidence the current Android/iOS app source (not just legacy PAM/CLI code) is published, and any independent corroboration of license terms.

   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [github] “The Google Authenticator project includes implementations of one-time passcode generators for several mobile platforms.”
   • [github] “Pluggable Authentication Module, aka PAM.”

   proof ↗⚑ flag

   2FAS✓fullprobed8/10

   2FAS explicitly points to its GitHub repositories for both the auth app and browser extension, describing itself as open-source, and the probe confirms 'open-source two-factor authentication' positioning with 'Freedom to self-host your data.' missing for 10: explicit license name (e.g., MIT/GPL) is not stated in evidence, and no independent audit confirming license terms.

   • [claimed-docs] “You can easily take a look at our code - reflecting our transparent and user-first approach. Check out our GitHub repository”
   • [claimed-docs] “Join our GitHub project and contribute to 2FAS by running code review and committing your ideas.”
   • [claimed-docs] “Check out our GitHub repository and delve deep into the intricacies of the browser extension back-end.”
   • [claimed-docs] “Freedom to self-host your data”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

4. ai-native userSelf-host the core product

   weight 3 · round to 2FAS
   Google Authenticator—none0/10

   Google Authenticator's core mobile app is closed-source and tied to Google account infrastructure for sync (docs-1, docs-4); the only open-source artifacts on GitHub are OTP-generator libraries and a PAM module, not the actual app users run, and there's no documented way to self-host the core product. Community threads reinforce this, complaining about lack of control over codes and dependency on Google's cloud (comm-6, comm-7, comm-12).

   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [github] “Pluggable Authentication Module, aka PAM.”
   • [community] “Storing it in the google cloud doesn't satisfy me. I just simply want the codes under my control... google still makes it stupendously diffi…”
   • [community] “If Google ever decides to kick you out of your account, Authenticator data will be gone. Google has done this on several occasions in the pa…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS~partialcommunity6/10

   2FAS explicitly markets 'Freedom to self-host your data' and lets users choose their own cloud/WebDAV/external drive storage for vaults, and the core apps are open-source on GitHub for self-hosting/auditing. However, this is self-hosting of data/storage, not the full core service (no self-hostable server/backend, no API for self-managed infrastructure), and community commentary questions how this differs meaningfully from server-reliant competitors. Missing for 10: a self-hostable backend/server component, documentation of running the whole product on own infrastructure, and independent confirmation that self-hosted storage removes all vendor dependency.

   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [claimed-docs] “Freedom to self-host your data”
   • [claimed-docs] “You can easily take a look at our code - reflecting our transparent and user-first approach. Check out our GitHub repository”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

Storage

1. everyday userThe app stores passkeys and signs me into websites and apps with them, not just TOTP codes

   weight 3 · round drawn
   Google Authenticator—none0/10

   All evidence describes Google Authenticator solely as a TOTP/HOTP code generator with cloud sync of codes; there is no mention of passkey storage or passkey-based sign-in anywhere in the docs, GitHub description, or community discussion.

   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”

   proof ↗⚑ flag

   2FAS—none0/10

   2FAS is a TOTP authenticator (2FAS Auth) and separately a password manager (2FAS Pass); evidence covers token/2FA codes, vault items, and password autofill, but nothing describes passkey (WebAuthn/FIDO2) storage or passkey-based sign-in to websites/apps.

   ⚑ flag

Sync

1. power userMy passkeys sync across ecosystems — iOS, Android, Windows, Linux, browsers — instead of being locked to one platform vendor

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   The evidence pack describes 2FAS Auth's 2FA token sync and 2FAS Pass's password vault sync (iCloud, Google Drive, WebDAV, browser extension) but contains no mention of 'passkey' or WebAuthn/FIDO2 credential support or storage anywhere. Since passkey sync is a distinct capability from OTP tokens or password items, there's no evidence this product delivers it.

   ⚑ flag

Export

1. power userExport all my TOTP secrets in an open, readable format and leave for another app whenever I choose — no lock-in

   weight 3 · round to 2FAS
   Google Authenticator!disputedcontradicted3/10

   Docs and community posts confirm a QR-code 'export' exists, but it is a proprietary migration QR built for transferring to another Google Authenticator instance, not an open/readable text format, and users explicitly report Google 'makes it stupendously difficult to just get a simple text export to a file' and that codes cannot be included in backups. This directly contradicts the story's requirement of an open, readable, portable export for switching apps freely. Missing for 10: an official plain-text/otpauth-URI export, documented compatibility with third-party apps, and confirmation the QR format is standardized/open rather than proprietary.

   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “To manually transfer Authenticator codes to a new device, you need: Your old device with Google Authenticator codes”
   • [community] “Google Authenticator already has a QR-Code based very easy export procedure, I just backup my GAuth to my spare phone and tablet. It feels s…”
   • [community] “Google Authenticator has a feature to allow export of all saved codes into one QR code for re-import into Google Authenticator running on an…”
   • [community] “Storing it in the google cloud doesn't satisfy me. I just simply want the codes under my control... google still makes it stupendously diffi…”
   • [community] “Google Authenticator does not allow you to include your codes in your backup, even when it's encrypted. So due to Google Authenticator, losi…”

   proof ↗⚑ flag

   2FAS~partialclaimed6/10

   2FAS explicitly supports imports from other apps (Aegis, Raivo, Google Authenticator, etc.) and offers 2FAS export files for backup/sync, with optional password-based encryption of backups. However, evidence doesn't confirm the export format is an open/readable standard (e.g., plaintext otpauth URIs or documented JSON schema) rather than a proprietary encrypted blob, and there's no explicit statement about exporting TO other apps' formats for migration away from 2FAS. missing for 10: documented open/standard export format spec, explicit 'export to other apps' migration path, independent confirmation the export file is human-readable/portable.

   • [claimed-docs] “We support imports from Aegis, Raivo, LastPassAegis, Stratum (formerly Authenticator Pro) and Google Authenticator.”
   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”

   proof ↗⚑ flag

Import

1. power userImport tokens in bulk from other authenticator apps instead of re-enrolling every account by hand

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   Evidence only covers transferring/syncing Google Authenticator codes between devices running Google Authenticator itself (QR export/import, Google Account sync), not bulk import of tokens originating in other authenticator apps (e.g., Authy, Microsoft Authenticator). No documentation or community report describes cross-app bulk import.

   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [claimed-docs] “To manually transfer Authenticator codes to a new device, you need: Your old device with Google Authenticator codes”
   • [community] “Google Authenticator has a feature to allow export of all saved codes into one QR code for re-import into Google Authenticator running on an…”

   proof ↗⚑ flag

   2FAS~partialclaimed6/10

   2FAS docs explicitly list bulk import support from Aegis, Raivo, LastPassAegis, Stratum, and Google Authenticator, directly supporting the power-user portability story. However, the list is limited to a handful of named apps, no evidence of a generic/universal import format (e.g., OTP URI batch, QR export from arbitrary apps), and no independent hands-on verification of the import process working smoothly. Missing for 10: broader import coverage beyond five named apps, independent/community confirmation of successful bulk import, and details on edge cases (e.g., failed imports, partial data loss).

   • [claimed-docs] “We support imports from Aegis, Raivo, LastPassAegis, Stratum (formerly Authenticator Pro) and Google Authenticator.”

   proof ↗⚑ flag

Migration

1. everyday userMoving to a new phone transfers all my tokens in one guided step

   weight 2 · round to 2FAS
   Google Authenticator!disputedcontradicted4/10

   Google's own docs describe a guided one-step transfer via signing into a Google Account (auto-sync) or scanning a QR export code between old/new devices, which matches the story closely. However, multiple hands-on community reports concretely contradict this working reliably in practice—users report app upgrades wiping all tokens, phone replacements losing TOTP keys despite expecting sync/backup, and account bans erasing synced codes entirely. missing for 10: independent verification that the guided transfer works consistently without data loss, and resolution of documented upgrade/sync failures.

   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [community] “I recently had a broken phone replaced and had depended on a backup to have my TOTP keys on my new phone. It was not a part of the phone bac…”
   • [community] “Don't upgrade! I just had this unpleasant experience and warned everyone... Warning: Google Authenticator upgrade loses all accounts.”
   • [community] “Every time I upgraded to a new iOS 7 beta, it wiped my Google Authenticator account tokens... I can't log in to my CampBX account anymore.”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS✓fullclaimed7/10

   2FAS Auth supports iCloud/Google Drive sync and encrypted export/backup files, letting a user restore tokens on a new device by simply enabling sync or importing the backup file, and 2FAS Pass has a dedicated recovery flow for restoring the vault on a different device via iCloud/Google Drive backups. missing for 10: a documented single-tap 'migrate to new phone' wizard UI and independent hands-on confirmation of the transfer experience.

   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [claimed-docs] “Multi-device sync (encrypted)”
   • [claimed-docs] “Learn how to restore access to your 2FAS Pass Vault or run it on a different device.”
   • [claimed-docs] “If you have enabled sync with iCloud (on iOS) or Google Drive (on Android), your encrypted 2FAS Pass Vault file is right there.”

   proof ↗⚑ flag

1. ai-native userChoose where my data is stored (region/residency)

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS~partialcommunity7/10

   2FAS is local-first: users choose to keep data on-device, in their own iCloud/Google Drive, WebDAV, or fully self-hosted, giving strong control over storage location (2fas-docs-5, 2fas-docs-14, 2fas-docs-17). This satisfies the spirit of data residency choice, though it's not a formal 'select region/datacenter' feature and relies on user-configured third-party clouds rather than a first-party regional hosting option. Missing for 10: explicit region/datacenter selection controls, independent verification of self-host setup beyond marketing pages.

   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [claimed-docs] “Freedom to self-host your data”
   • [claimed-docs] “Multi-device sync (encrypted)”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

2. ai-native userPrevent my data from being used to train AI models

   weight 3 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS~partialprobed4/10

   2FAS is local-first with explicit 'no telemetry' claims and no public API for data collection (2fas-probe-rt-1, 2fas-probe-rt-2), which implies user data isn't centrally harvested and thus couldn't feed AI training pipelines. However, there is no explicit statement addressing AI model training or an opt-out policy specifically. Missing for 10: explicit AI-training data policy/opt-out statement, independent confirmation that no data is ever sent to third-party AI services.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE runtime (recorded 2026-09-15): machine-readable pricing verified live at https://2fas.com/.well-known/pricing.md — '# 2FAS Pricing... …”

   proof ↗⚑ flag

3. ai-native userControl data retention and deletion

   weight 2 · round to 2FAS
   Google Authenticator~partialcommunity4/10

   Google Authenticator lets users delete individual codes (swipe to delete), edit/reassign codes, and use the app without a Google account (avoiding cloud retention entirely), giving some control over what data is stored and removed. However, there is no documented account-level 'delete all my data' or data-export-to-file control, and community reports describe difficulty exporting/backing up codes and total data loss if a Google account is banned or deleted, undermining full user control over retention. missing for 10: explicit data export/portability controls, account-wide deletion/retention settings, and confirmation that deleting codes removes them from Google's backend, not just the local view.

   • [claimed-docs] “To delete an Authenticator code on Android, swipe right on any code to show the delete option.”
   • [claimed-docs] “When you first open the app, tap Use without an account.”
   • [claimed-docs] “For additional protection, you can turn on Privacy Screen in Google Authenticator. This requires a verification from your device, like a PIN…”
   • [community] “Storing it in the google cloud doesn't satisfy me. I just simply want the codes under my control... google still makes it stupendously diffi…”
   • [community] “Google Authenticator does not allow you to include your codes in your backup, even when it's encrypted. So due to Google Authenticator, losi…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS~partialprobed6/10

   2FAS gives users explicit control over where data is stored (own cloud, external drive, self-hosting) and confirms no telemetry/tracking, which supports retention control (2fas-docs-5, 2fas-docs-14, 2fas-probe-rt-2). However there is no explicit documentation of a deletion/erasure workflow, data-retention policy, or account-deletion mechanism beyond implied local file control. Missing for 10: explicit data deletion/right-to-erasure documentation, retention period policy, and independent confirmation that deleting local/cloud copies fully removes data.

   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [claimed-docs] “Freedom to self-host your data”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [probe] “PROBE runtime (recorded 2026-09-15): machine-readable pricing verified live at https://2fas.com/.well-known/pricing.md — '# 2FAS Pricing... …”

   proof ↗⚑ flag

4. ai-native userOpt out of telemetry and usage tracking

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS✓fullprobed8/10

   2FAS explicitly states in its machine-readable pricing/docs that 2FAS Auth has 'no advertising, no telemetry,' meaning tracking is absent by design rather than requiring opt-out, and the product is positioned as local-first/open-source, reinforcing a privacy-first posture. Missing for 10: no explicit dedicated privacy-settings toggle documentation, and no independent audit confirming zero telemetry across 2FAS Pass as well as Auth.

   • [probe] “PROBE runtime (recorded 2026-09-15): machine-readable pricing verified live at https://2fas.com/.well-known/pricing.md — '# 2FAS Pricing... …”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

App lock

1. everyday userThe app itself locks behind biometrics or a PIN, so a borrowed phone doesn't expose my codes

   weight 2 · round to Google Authenticator
   Google Authenticator✓fullclaimed8/10

   Google's own docs describe a 'Privacy Screen' feature requiring PIN, pattern, or biometric verification before the app can be used, directly matching the story. No community evidence contradicts this specific lock-screen feature (complaints focus on sync/backup, not app-lock). Missing for 10: independent hands-on confirmation of the Privacy Screen feature working reliably across devices.

   • [claimed-docs] “For additional protection, you can turn on Privacy Screen in Google Authenticator. This requires a verification from your device, like a PIN…”

   proof ↗⚑ flag

   2FAS—none0/10

   The evidence pack describes 2FAS Pass's tiered security and sync options, and confirms access to the browser extension/autofill must be approved in-app, but nowhere states that the 2FAS Auth or Pass app itself can be locked behind biometrics or a PIN to block access if the phone is borrowed. No docs or community evidence mention an app-lock/biometric-gate feature for the authenticator itself.

   • [claimed-docs] “2FAS Pass supports 3 Security Tiers for protecting your Items.”
   • [claimed-docs] “Access through the Browser Extension or using Autofill must be confirmed in 2FAS Pass.”

   proof ↗⚑ flag

Audits

1. security engineerThe vendor publishes independent security audits of the app and its sync protocol

   weight 2 · round drawn
   Google Authenticator—none0/10

   No evidence of any independent security audit of the app or its cloud sync protocol; a notable 2023 independent security researcher analysis found the sync data was unencrypted, but no formal published audit exists in the evidence pack, only community complaints about sync/backup reliability and trust concerns.

   ⚑ flag

   2FAS—none0/10

   Evidence shows only open-source code availability claims and general transparency messaging, with no mention of independent third-party security audits of the app or its sync protocol; community discussion even questions the trust model of cloud-based sync without referencing any audit report.

   • [claimed-docs] “You can easily take a look at our code - reflecting our transparent and user-first approach. Check out our GitHub repository”
   • [claimed-docs] “Join our GitHub project and contribute to 2FAS by running code review and committing your ideas.”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

Monitoring

1. power userThe app warns me when a service I use is breached or a stored credential is weak, reused, or exposed

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   No evidence of breach monitoring, weak/reused/exposed password detection, or security-posture alerts anywhere in the docs, community reports, or probes; 2FAS's evidence covers sync, tiers, self-hosting, and import/export but nothing about credential health warnings or breach alerts.

   ⚑ flag

Push

1. it adminPush-based sign-in includes phishing defenses — number matching, location context, and admin-enforced MFA policies

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

Server

1. it adminSelf-host the sync server on my own infrastructure and keep every secret inside my perimeter

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS~partialcommunity4/10

   2FAS advertises 'freedom to self-host your data' and lets users choose where encrypted vault/token data lives (own cloud, WebDAV, external drive) rather than relying on 2FAS-hosted storage, which partially satisfies the perimeter-control need. However there is no evidence of an actual downloadable/self-hostable 'sync server' component (e.g. a Vaultwarden-style server binary) — it's more bring-your-own-storage for encrypted files/backups, and community commentary explicitly questions whether this differs meaningfully from relying on third-party server infrastructure. Missing for 10: a dedicated self-hosted sync-server artifact/deployment docs, and independent confirmation that all sync traffic stays fully within admin-controlled infra.

   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [claimed-docs] “Freedom to self-host your data”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

Source

1. security engineerThe client apps are open source, so the code handling my seeds can be inspected and community-reviewed

   weight 2 · round to 2FAS
   Google Authenticator~partialclaimed4/10

   The GitHub repo (google/google-authenticator) documents open-source implementations of HOTP/TOTP and a PAM module, giving some basis for code review, but the evidence pack has no confirmation that the current shipped Android/iOS apps (which now handle cloud sync of seeds) actually correspond to this repo or receive active community review. missing for 10: evidence that the live mobile app source matches the published repo, evidence of active maintenance/community audits of the seed-handling code, and independent confirmation of code correctness.

   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [github] “The Google Authenticator project includes implementations of one-time passcode generators for several mobile platforms.”
   • [github] “Pluggable Authentication Module, aka PAM.”

   proof ↗⚑ flag

   2FAS✓fullprobed8/10

   2FAS explicitly markets itself as open-source and points to a public GitHub repo for code review, including for the browser extension backend ('delve deep into the intricacies'), and invites community contribution/code review. The core apps (2FAS Auth, described as 'Free, Open-source 2FA Authenticator App') are confirmed open-source via probe data. Missing for 10: no independent security audit report or third-party code-review corroboration is cited, and evidence doesn't explicitly confirm the mobile app repos (vs. just extension) are inspectable.

   • [claimed-docs] “You can easily take a look at our code - reflecting our transparent and user-first approach. Check out our GitHub repository”
   • [claimed-docs] “Join our GitHub project and contribute to 2FAS by running code review and committing your ideas.”
   • [claimed-docs] “Check out our GitHub repository and delve deep into the intricacies of the browser extension back-end.”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

Automation

1. ai-native userAn agent can fetch a current TOTP code programmatically — via CLI or API — to complete a 2FA login inside an automated workflow

   weight 2 · round drawn
   Google Authenticator—none0/10

   Google Authenticator is a mobile/PAM app with no documented CLI or API for programmatic TOTP retrieval; evidence only covers manual app UI actions (sync, QR export, editing codes) and community complaints about backup/sync issues, none of which show agent-accessible code fetching.

   ⚑ flag

   2FAS—none0/10

   The evidence explicitly states 2FAS has no public API for programmatic TOTP retrieval ('Public API: None — local-first by design, no agent-callable endpoint for token generation'), and no CLI is documented anywhere in the pack; OpenAPI probes all 404. This directly rules out agent-driven fetching of TOTP codes via CLI or API.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

2. ai-native userExpose the vault through a local programmatic endpoint an agent can query for codes and secrets without screen-scraping the app

   weight 1 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   The product's own llms.txt explicitly states 'Public API: None' and no OpenAPI/local endpoint was found; 2FAS is local-first by design with no agent-callable interface for querying codes or secrets programmatically.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

Browser

1. everyday userA browser extension autofills my TOTP codes and passkeys during login instead of making me retype them

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS✓fullcommunity7/10

   2FAS's browser extension is documented to autofill TOTP tokens with confirmation via the mobile app ('You don't have to retype anything'), and the 2FAS Pass extension supports autofill for passwords/passkeys confirmed in the vault app, with the free tier confirmed by community reports to include auto-fill. However, evidence is vendor-documented with only partial independent corroboration and no explicit hands-on confirmation of passkey autofill specifically working smoothly in-browser. Missing for 10: independent hands-on validation of passkey autofill UX, more detail on cross-browser support, and confirmation of friction-free flow beyond vendor claims.

   • [claimed-docs] “Click the browser extension icon when you need to enter a 2FA token, then confirm in the mobile app. You don't have to retype anything.”
   • [claimed-docs] “Access through the Browser Extension or using Autofill must be confirmed in 2FAS Pass.”
   • [community] “Free tier of 2FAS Pass includes auto-fill, import/export, and standard password generator; paid tier adds multi-device sync, more than 200 v…”
   • [claimed-docs] “You can easily take a look at our code - reflecting our transparent and user-first approach. Check out our GitHub repository”

   proof ↗⚑ flag

Desktop

1. power userA first-class desktop app gives me codes and vault access on my computer, not just on mobile

   weight 2 · round drawn
   Google Authenticator—none0/10

   No evidence of a desktop application; Google Authenticator is documented and discussed only as a mobile app with QR-based transfer/sync between phones, not a desktop client or vault.

   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [github] “The Google Authenticator project includes implementations of one-time passcode generators for several mobile platforms.”

   proof ↗⚑ flag

   2FAS—none0/10

   Evidence only shows mobile apps (iOS/Android) plus a browser extension, and the extension explicitly requires confirmation via the mobile app (2fas-docs-3, 2fas-docs-19) rather than functioning as an independent desktop client. No native desktop app for code generation or vault access is documented anywhere in the pack.

   • [claimed-docs] “Click the browser extension icon when you need to enter a 2FA token, then confirm in the mobile app. You don't have to retype anything.”
   • [claimed-docs] “Access through the Browser Extension or using Autofill must be confirmed in 2FAS Pass.”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

Watch

1. everyday userRead my codes from my smartwatch without pulling out my phone

   weight 1 · round to 2FAS
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS~partialclaimed5/10

   2FAS explicitly markets 'Apple Watch — Improved convenience and accessibility' as a feature, indicating watch support exists, but there's no detail on what functionality is available on the watch (e.g., whether codes are viewable standalone or require phone proximity) and no independent/hands-on confirmation of the experience. missing for 10: details on watch app functionality/UX, confirmation codes can be read without unlocking phone, independent hands-on review of the watch experience.

   • [claimed-docs] “Apple Watch Improved convenience and accessibility.”
   • [claimed-docs] “Apple WatchImproved convenience and accessibility.”

   proof ↗⚑ flag

Backup

1. security engineerBackups are end-to-end encrypted with a key the vendor never holds, and the encryption design is documented

   weight 3 · round to 2FAS
   Google Authenticator—none0/10

   Docs only describe syncing codes via a Google Account (google-authenticator-docs-1,4) with no mention of end-to-end encryption or a vendor-blind key, and community reports explicitly note that storing codes in 'google cloud doesn't satisfy me' and that losing access to the Google account means losing all 2FA data, implying no such E2E encryption exists.

   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [community] “Storing it in the google cloud doesn't satisfy me. I just simply want the codes under my control... google still makes it stupendously diffi…”
   • [community] “If Google ever decides to kick you out of your account, Authenticator data will be gone. Google has done this on several occasions in the pa…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS~partialcommunity4/10

   2FAS documents that backups/sync can be encrypted with a user password (docs-9, docs-17, docs-18) and that vault storage location is user-controlled, but the encryption is presented as optional ('we encourage you to do that') rather than guaranteed end-to-end by default, and there is no detailed technical whitepaper describing key derivation, zero-knowledge architecture, or threat model. Community discussion (2fas-comm-1) explicitly questions whether backups to iCloud/WebDAV truly avoid vendor/server trust, further undercutting a strong E2E claim. missing for 10: a dedicated cryptographic design document (algorithms, key derivation, zero-knowledge proof), confirmation that encryption is mandatory/default rather than opt-in, and independent security audit corroboration.

   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [claimed-docs] “Multi-device sync (encrypted)”
   • [claimed-docs] “If you have enabled sync with iCloud (on iOS) or Google Drive (on Android), your encrypted 2FAS Pass Vault file is right there.”
   • [claimed-docs] “You decide where to store your Items. You can store it in your own cloud or keep it as an encrypted file on an external drive.”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

Multi device

1. everyday userMy tokens are available on my phone, tablet, and computer at the same time, kept in sync automatically

   weight 2 · round to 2FAS
   Google Authenticator~partialcommunity5/10

   Google's docs confirm automatic cross-device sync via Google Account sign-in, works offline, and supports multiple accounts (docs-1,4,5,10), but Authenticator is a mobile-only app with no native desktop/computer client, so the 'computer' leg of the story is unsupported. Community reports also cite historical sync/backup failures and account-loss risk tying codes to a Google Account (comm-1, comm-7, comm-12, comm-13), showing real caveats even if not disproving the current sync feature outright. Missing for 10: a computer/desktop client, independent hands-on confirmation the modern sync feature works reliably across all device types without data loss.

   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [claimed-docs] “Authenticator can sync codes for multiple Google Accounts and display them from the same mobile device.”
   • [claimed-docs] “You can still generate codes without an internet connection or mobile service.”
   • [community] “I recently had a broken phone replaced and had depended on a backup to have my TOTP keys on my new phone. It was not a part of the phone bac…”
   • [community] “If Google ever decides to kick you out of your account, Authenticator data will be gone. Google has done this on several occasions in the pa…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”
   • [community] “Don't upgrade! I just had this unpleasant experience and warned everyone... Warning: Google Authenticator upgrade loses all accounts.”

   proof ↗⚑ flag

   2FAS~partialcommunity6/10

   2FAS documents syncing tokens across devices via iCloud/Google Drive and lists 'Multi-device sync (encrypted)' as a feature, and computer access is possible through a browser extension that confirms via the mobile app. However, this cross-device access relies on manual confirmation from the phone rather than tokens being independently and automatically available on a tablet/computer at all times, and multi-device sync appears gated behind a paid tier for the Pass product. Missing for 10: evidence of true simultaneous, always-available token access across phone/tablet/computer without per-use phone confirmation, and confirmation that Auth (not just Pass) multi-device sync is free and automatic.

   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “Click the browser extension icon when you need to enter a 2FA token, then confirm in the mobile app. You don't have to retype anything.”
   • [claimed-docs] “Multi-device sync (encrypted)”
   • [community] “Free tier of 2FAS Pass includes auto-fill, import/export, and standard password generator; paid tier adds multi-device sync, more than 200 v…”

   proof ↗⚑ flag

Recovery

1. everyday userLosing my phone doesn't lose my accounts — a documented recovery path restores my tokens on a new device

   weight 3 · round to 2FAS
   Google Authenticator!disputedcontradicted5/10

   Google's own docs describe two documented recovery paths — signing into your Google Account to auto-sync codes across devices, or scanning a QR-code export to transfer codes to a new device (google-authenticator-docs-1/3/4/11) — which is exactly the story's ask. However, multiple hands-on community reports concretely contradict reliability of this recovery: users report app upgrades wiping all tokens with no recovery (comm-13, comm-14), phone backups explicitly excluding Authenticator data (comm-1, comm-8), and losing all codes permanently if the tied Google Account is banned/deleted (comm-7, comm-12). Missing for 10: independent verification that account-sync recovery works consistently across real device-loss scenarios, and resolution of the backup-exclusion/account-ban failure modes.

   • [claimed-docs] “With Google Authenticator, you can synchronize your verification codes across all your devices, simply by signing in to your Google Account.”
   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “When you sign in to your Google Account within Google Authenticator on a new device, your codes are automatically synced to this device.”
   • [claimed-docs] “To manually transfer Authenticator codes to a new device, you need: Your old device with Google Authenticator codes”
   • [community] “I recently had a broken phone replaced and had depended on a backup to have my TOTP keys on my new phone. It was not a part of the phone bac…”
   • [community] “Does that mean one can use adb to backup Google Authenticator's data as well? Last time I tried, the app data was explicitly marked as exclu…”
   • [community] “Don't upgrade! I just had this unpleasant experience and warned everyone... Warning: Google Authenticator upgrade loses all accounts.”
   • [community] “Every time I upgraded to a new iOS 7 beta, it wiped my Google Authenticator account tokens... I can't log in to my CampBX account anymore.”
   • [community] “If Google ever decides to kick you out of your account, Authenticator data will be gone. Google has done this on several occasions in the pa…”
   • [community] “Imagine your google account getting deleted cuz you got banned from Google and the suddenly you lose all your 2FA secrets cuz they are part …”

   proof ↗⚑ flag

   2FAS✓fullcommunity8/10

   2FAS documents multi-device sync via iCloud/Google Drive/encrypted export files for the Auth app, plus a dedicated recovery page explaining how to restore or run the Pass Vault on a new device using the cloud-synced encrypted vault file, and supports importing from other authenticator apps for migration scenarios. This directly addresses losing a phone and restoring tokens elsewhere, though community commentary raises trust questions about relying on third-party cloud storage rather than disputing that recovery works. Missing for 10: independent hands-on verification of the actual recovery flow succeeding, and clearer detail on recovery when no cloud sync was enabled beforehand.

   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “Learn how to restore access to your 2FAS Pass Vault or run it on a different device.”
   • [claimed-docs] “If you have enabled sync with iCloud (on iOS) or Google Drive (on Android), your encrypted 2FAS Pass Vault file is right there.”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”
   • [claimed-docs] “We support imports from Aegis, Raivo, LastPassAegis, Stratum (formerly Authenticator Pro) and Google Authenticator.”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”

   proof ↗⚑ flag

Automation

1. ai-native userAn agent can create and update vault entries — seeding new TOTP secrets, storing credentials — through documented programmatic surfaces

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   No programmatic API exists for creating/updating vault entries or seeding TOTP secrets; the product's own llms.txt explicitly states 'Public API: None' and openapi probes all 404, confirming no agent-callable surface for this workflow.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://2fas.com/openapi.json, https://2fas.com/swagger.json, https://2fas.com/api/openapi.json, htt…”

   proof ↗⚑ flag

Policy

1. it adminEnforce org-wide policies — require app lock, restrict export, mandate strong master credentials — across every member's app

   weight 1 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is a consumer-focused authenticator/password manager with no evidence of centralized org policy management, admin dashboards, enforced app-lock policies, export restrictions, or credential strength mandates across users; it is explicitly local-first with no public API for such control (2fas-probe-rt-1). No IT-admin console or MDM/policy feature is documented anywhere in the pack.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [claimed-docs] “Freedom to self-host your data”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”

   proof ↗⚑ flag

Sharing

1. it adminMy team can share TOTP-protected logins through shared vaults or collections with per-member access control

   weight 2 · round drawn
   Google Authenticator—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

   2FAS—none0/10

   2FAS is positioned as a personal, local-first authenticator/password manager with individual sync (iCloud/Google Drive/self-host) and multi-device sync for a single user, not team/organization features. No evidence of shared vaults, collections, or per-member access control for an IT-admin-managed team; the HN community discussion even questions its single-user 'local-only' security model, but nothing addresses team sharing at all.

   • [claimed-docs] “Synchronize your Tokens with 2FAS export files, iCloud or Google Drive.”
   • [claimed-docs] “Multi-device sync (encrypted)”
   • [community] “Skeptical about 'local-only' claim: 2FAS Pass supports backups to iCloud, WebDAV and local file, meaning passwords rely on a server's securi…”
   • [community] “Free tier of 2FAS Pass includes auto-fill, import/export, and standard password generator; paid tier adds multi-device sync, more than 200 v…”

   proof ↗⚑ flag

Enrollment

1. everyday userScan a QR code and immediately start generating TOTP codes for a new account

   weight 3 · round to Google Authenticator
   Google Authenticator✓fullclaimed8/10

   Docs and GitHub evidence confirm the app scans QR codes to add accounts and generates RFC 6238 TOTP codes, including offline (docs-3, docs-10, gh-1). Community feedback criticizes backup/sync/loss issues, not the basic scan-and-generate flow, so the core story is undisputed. Missing for 10: independent hands-on confirmation of the exact scan-to-first-code experience and no first-party UX walkthrough beyond support docs.

   • [claimed-docs] “On your old device, create a QR code... On your new device, tap Scan QR code”
   • [claimed-docs] “You can still generate codes without an internet connection or mobile service.”
   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [github] “The Google Authenticator project includes implementations of one-time passcode generators for several mobile platforms.”

   proof ↗⚑ flag

   2FAS~partialprobed5/10

   2FAS Auth is a well-known TOTP app and docs reference connecting accounts via video tutorials, but the evidence pack lacks any direct documentation or hands-on confirmation of the core QR-scan-to-code-generation flow. missing for 10: explicit documentation/screenshots of QR scanning UX, confirmation that codes generate immediately post-scan, independent/hands-on corroboration of the onboarding flow.

   • [claimed-docs] “use our quick and easy-to-follow video tutorials to connect your online accounts.”
   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”

   proof ↗⚑ flag

Offline

1. power userCodes generate fully offline — no network, no vendor account required just to see my TOTP codes

   weight 2 · round drawn
   Google Authenticator✓fullclaimed8/10

   Docs explicitly state you can 'Use without an account' at setup and that codes generate 'without an internet connection or mobile service,' confirming fully offline, account-free TOTP generation (core RFC 6238 implementation per GitHub). Community discussion focuses on backup/sync/account-loss risks, not on offline generation failing, so no contradiction of this specific axis. Missing for 10: independent hands-on confirmation specifically testing offline-without-account code generation.

   • [claimed-docs] “When you first open the app, tap Use without an account.”
   • [claimed-docs] “You can still generate codes without an internet connection or mobile service.”
   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”

   proof ↗⚑ flag

   2FAS✓fullprobed8/10

   2FAS Auth is described as open-source, local-first with no public API for token generation, and pricing docs confirm it's free with no telemetry, consistent with fully offline TOTP generation without vendor account requirement. Community evidence questions cloud-sync/backup security but this concerns 2FAS Pass (password manager), not the core Auth TOTP generation flow. missing for 10: independent hands-on confirmation that TOTP codes generate with device fully offline (airplane mode test), and explicit documentation stating no account/signup is ever required.

   • [probe] “PROBE runtime (recorded 2026-09-15): 2fas.com serves the most agent-forward docs plumbing in the arena — llms.txt with a literal 'For AI age…”
   • [probe] “PROBE runtime (recorded 2026-09-15): machine-readable pricing verified live at https://2fas.com/.well-known/pricing.md — '# 2FAS Pricing... …”
   • [claimed-docs] “data can be encrypted with your password, so we encourage you to do that.”

   proof ↗⚑ flag

Organization

1. power userKeep dozens of tokens organized — search, folders or groups, and service icons — so the right code is always two taps away

   weight 2 · round to 2FAS
   Google Authenticator—none0/10

   Evidence only shows basic manual reordering (drag-to-reorder), swipe-to-edit/delete, and username/account editing — there is no mention of search, folders/groups, or service icons for organizing many tokens. missing for 10: search functionality, folder/group organization, service icons, any at-scale organization feature for dozens of tokens.

   • [claimed-docs] “To edit your Authenticator code on Android, swipe left on any code to show the edit option.”
   • [claimed-docs] “To organize your Authenticator codes, touch and hold any code, then drag to reorder to a desired location.”
   • [claimed-docs] “To delete an Authenticator code on Android, swipe right on any code to show the delete option.”
   • [claimed-docs] “To edit your Authenticator code on Android, swipe left on any code to show the edit option. You can update the username for the code or chan…”

   proof ↗⚑ flag

   2FAS~partialclaimed6/10

   Docs confirm token customization with 'full control over groups, badges, icons, and labels' and Apple Watch access, directly supporting organization by groups/icons for quick access, but there is no mention of search functionality within the token list, no detail on folder/group UX depth, and no independent hands-on account of managing 'dozens of tokens' at scale. missing for 10: explicit search feature evidence, hands-on/independent validation of organization UX at scale, deeper detail on folder vs group implementation.

   • [claimed-docs] “Token customization Full control over the groups, badges, icons, and labels.”
   • [claimed-docs] “Token customizationFull control over the groups, badges, icons, and labels.”
   • [claimed-docs] “Apple Watch Improved convenience and accessibility.”
   • [claimed-docs] “Apple WatchImproved convenience and accessibility.”

   proof ↗⚑ flag

Variants

1. power userNon-standard tokens work too — Steam Guard, HOTP counters, custom periods and digit lengths

   weight 2 · round to Google Authenticator
   Google Authenticator~partialclaimed3/10

   The GitHub docs confirm the underlying library supports HOTP (RFC 4226) and TOTP (RFC 6238) algorithms, giving some basis for HOTP counter support, but there is no evidence anywhere in the pack that the consumer app exposes Steam Guard mode, custom periods, or custom digit lengths — the mobile app UI evidence only shows scanning QR codes, editing username, and deleting entries. missing for 10: Steam Guard token support, custom period configuration, custom digit-length configuration, and any UI/setting for manual HOTP counter entry.

   • [github] “These implementations support the HMAC-Based One-time Password (HOTP) algorithm specified in RFC 4226 and the Time-based One-time Password (…”
   • [github] “The Google Authenticator project includes implementations of one-time passcode generators for several mobile platforms.”
   • [claimed-docs] “To edit your Authenticator code on Android, swipe left on any code to show the edit option.”

   proof ↗⚑ flag

   2FAS—none0/10

   No evidence in the pack addresses Steam Guard tokens, HOTP counter-based codes, or custom period/digit-length configuration for 2FAS Auth — the docs only mention generic 'token customization' for groups/badges/icons/labels, which is unrelated to non-standard token algorithms.

   • [claimed-docs] “Token customization Full control over the groups, badges, icons, and labels.”
   • [claimed-docs] “Token customizationFull control over the groups, badges, icons, and labels.”

   proof ↗⚑ flag