Your AI chats.One local library.

Folders, search, cleanup, and export for the AI apps you already use — everything stays on your device.

No Tecora account · No backend · Chrome MV3

Not an AI wrapper. A local layer that makes living across Claude, ChatGPT, and Gemini less chaotic.

τ TecoraAll platformsSearch chats…AllshipresearchpersonalAuth middleware for edge routesuser → assistant · claudeRefactor pricing calcuser → assistant · chatgptTrip itinerary — Lisbonuser → assistant · geminiSQL join on tenantsuser → assistant · claudeLanding hero copy passuser → assistant · chatgptτ
01 — Capture

Start on the sites you already use.

Tecora reads your chat lists from Claude, ChatGPT, and Gemini while you browse. No CSV ritual to get started.

claude.aichatgpt.comgemini.google.comτLocal libraryIndexedDB
02 — Organize

Folders, tags, and resume — across models.

Pin what matters. Continue where you left off. Switch from one account to All platforms when you need the whole picture.

τ TecoraAll platforms · 342 chatsFOLDERSAllUnassignedshipresearchpersonalTAGS#launch#bug#ideaCONTINUE WHERE YOU LEFT OFFAuth middleware for edge routesyou → assistant · claude · 12m agoLanding hero copy passyou → assistant · chatgpt · 2h agoTrip itinerary — Lisbonyou → assistant · gemini · yesterday
Export

Export that packs the work.

Markdown for reading. Portable JSON archives for restore. ZIP when you need files and artifacts — with an honest MISSING.md when something can't be fetched.

chats.md
Markdown · readable
export
tecora-archive.json
Portable · restore folders/tags
export
export.zip
Files, artifacts, images
export
MISSING.mdAssets we couldn't fetch are listed here — honest, not hidden.
Privacy

No account. No backend. Nothing leaves the device.

Keep message capture on for full-text search, or limit a platform to titles only. Wipe all Tecora data anytime.

Privacy
on device
claude.ai
Message capture on
chatgpt.com
Message capture on
gemini.google.com
Titles only
The product

Four surfaces. One quiet layer.

Browse your entire library, filter by folder or tag, and jump back into any chat.

  • Browse chats across supported platforms
  • Folders, tags, pins, select/export
  • Resume recent work
τ
Highlights

Small pieces, picked deliberately.

  • 01
    Local-first

    Your library lives in the browser — not on our servers. We don't run chat servers.

  • 02
    Cross-platform

    One resume list across Claude, ChatGPT, and Gemini.

  • 03
    Derived titles

    "New chat" becomes something you can actually recognize.

  • 04
    Previews

    One-line recall so you know the thread before you open it.

  • 05
    Bulk delete

    A safety-gated queue beats rage-clicking native modals.

  • 06
    Usage awareness

    Local activity estimates — clearly labeled, never fake quota.

  • 07
    Shadow DOM UI

    Palette and dock sit cleanly on top of host pages.

Install

Up in a minute.

No signup, no config. Just install and open the AI apps you already use.

DownloadChrome MV3 · Load unpacked for judges
  1. 01Install the Tecora extension
  2. 02Open Claude, ChatGPT, or Gemini while logged in
  3. 03Open the side panel — your library starts filling in
FAQ

Questions answered honestly.

No. Tecora doesn't answer prompts or wrap models. It organizes chats from Claude, ChatGPT, and Gemini.
Chat data Tecora stores stays in your browser's local storage (IndexedDB). There is no Tecora backend for syncing your chats.
No.
v0.1 targets Chromium via a Manifest V3 extension (Chrome-class browsers).
claude.ai, chatgpt.com, and gemini.google.com.
Yes. Per-platform toggles can keep capture to titles/metadata and remove stored message text.
Not today. Usage awareness is a local estimate from captured activity, labeled clearly.
Yes. Portable JSON archive export/import restores chats, messages, folders, and tags.
It's powerful and irreversible on the platform. Tecora uses a safety-gated queue and asks you to keep the site tab open.
Yes for the current hackathon / v0.1 distribution. No SaaS billing.
τ

Keep your AI history coherent.

Install Tecora and give every model the same local memory layer.

Free to try · Runs locally · No Tecora account
v0.1 · Hackathon build