← Open Source
sxwangsxwang1

chatpassport

Move AI conversations between ChatGPT, Claude, Gemini, and DeepSeek with a local-first, open conversation format—no servers, no tracking.

ApplicationsPersonalProductivityTypeScript
Open on GitHub
Momentum
+1stars in 24 hours+0.2%
625
Stars
5
Forks
+-3
This week
2
Contributors
Created 2026-08-26 · Updated 2026-10-15 · #6576 today
Top developers
sxwangsxwang1

sxwangsxwang1/chatpassport is ranked 6,576 with 625 stars.

Same-category comparison

sxwangsxwang1/chatpassport is ranked 6,576 among these 8 peers; anhaidgroup/deepmatcher is ranked 6,577.

Same-category comparison
NameRankMetrics
raojiacui/prompt-lens6,569
Language:
TypeScript
Stars:
627
Forks:
79
PKU-YuanGroup/OpenAI4S6,571
Language:
Python
Stars:
626
Forks:
72
Description:
Open-source AI agent for scientific research. Analyze data in Python/R with Claude, GPT, Gemini, and more.
xoai/sage-wiki6,572
Language:
Go
Stars:
626
Forks:
103
Description:
sage-wiki is a graph memory and knowledge base that AI agents and humans build and query together. Drop in documents; an LLM compiler turns them into an interlinked wiki with a knowledge graph. One Go binary scales it from a personal vault to a team hub to a company knowledge graph.
arsh342/careercompass6,574
Language:
TypeScript
Stars:
625
Forks:
134
Description:
A next-generation AI-powered career platform transforming how employers and job seekers connect.
anhaidgroup/deepmatcher6,577
Language:
Python
Stars:
624
Forks:
130
Description:
Python package for performing Entity and Text Matching using Deep Learning.
Haleclipse/CometixCode6,580
Language:
Rust
Stars:
623
Forks:
37
Description:
An unofficial Rust reimplementation of Anthropic's Claude Code terminal UI, built on iocraft.
checkmarble/marble6,596
Language:
HTML
Stars:
616
Forks:
90
Description:
Marble - the real time decision engine for fraud and AML.
slavabarkov/tidy6,609
Language:
Kotlin
Stars:
610
Forks:
48
Description:
Offline semantic Text-to-Image and Image-to-Image search on Android powered by quantized state-of-the-art vision-language pretrained CLIP model and ONNX Runtime inference engine.
Key parameters and specifications
Key parameters and specifications
ParameterValue
LanguageTypeScript
CategoryApplications
SubcategoryPersonal, Productivity
KeywordsClient integration, Context engineering, Memory management
Stars625
Forks5
Contributors2
Created date2026-08-26
Updated date2026-10-15
Stars in 1 day1
Stars in 7 days-3
Rank6,576
Previous rank15,716
Rank change9,140
7-day stars rank18,673
Previous 7-day stars rank18,611
7-day stars rank change-62
Total stars rank16,255
Previous total stars rank16,263
Total stars rank change8
README summary

Continue a conversation in ChatGPT, Claude, Gemini, or DeepSeek without rebuilding its context by hand. ChatPassport reads the conversation you choose, prepares a draft in another assistant, and leaves the final review and Send click to you. Get started · Capture modes · What transfers · Privacy · Build from source *The screenshots use demonstration conversations. ChatPassport works alongside the supported websites; it does not replace their accounts or interfaces. When a chat gets long, you hit a limit, or you want a second

assistant's perspective, copying the conversation message by message is tedious. ChatPassport carries the text and code it can read into a new draft, then lets you add your next question before anything is sent. ** Move between ChatGPT, Claude, Gemini, and DeepSeek. ** Use **Quick** for currently loaded messages or **Load history** to scroll for older messages. ** Review the message count, coverage notice, and estimated draft fit before continuing. ** If older messages must be left out, ChatPassport asks before inserting anything. You

review and send the result yourself. ** Export Markdown or , and import saved JSON later. ChatPassport runs without a ChatPassport account, API key, developer-operated server, analytics, or tracking. Get started The steps below describe ChatPassport **0. If your installed extension does not show Quick and Load history, check its version in ; Web Store updates can appear after review. Install ChatPassport from the Chrome Web Store. Open a conversation on a supported website and open the ChatPassport side panel. If the

website was already open during installation, refresh that tab once. Choose another assistant and a capture mode, then click **Prepare for [destination]**. Check the captured messages and coverage notice. Click **Continue in [destination]** to open the destination tab. Type your next question in the destination's normal editor. On the ChatPassport card, click **Add context to draft**. Confirm any proposed omissions, review the resulting draft, and use the website's own Send button. Once context is inserted into the destination editor, that

website can read or sync the draft even before you send it. It does not scroll or wait for history, and marks the result as partial. | You need the visible recent exchange quickly. | | **Load history** (default) | Scrolls to collect older rendered messages, then restores the page position. | While Load history is running, **Stop and use collected messages** keeps what has been found so far. Websites may hide, virtualize, collapse, or load history in ways the extension cannot fully observe; a

successful scan does not guarantee every past message was captured. Loading history may also cause the source website's normal network requests. After capture, **More options** lets you transfer all captured messages or only the latest 20, 50, or 100. It also contains JSON/Markdown export and JSON import. Exports contain the captured preview, regardless of the range selected for transfer. ChatPassport estimates how many complete recent messages fit before you continue, then recalculates using the question you actually typed. If

it must omit older messages, it shows the counts and asks you to confirm before changing the editor. It does

Frequently asked questions

How many stars and forks are recorded?

sxwangsxwang1/chatpassport has 625 stars and 5 forks.

What is the repository rank?

sxwangsxwang1/chatpassport is ranked 6,576.

What language and category are recorded?

sxwangsxwang1/chatpassport has 625 stars, uses TypeScript and belongs to Applications.

Which repository is closest in rank?

sxwangsxwang1/chatpassport is compared with anhaidgroup/deepmatcher, ranked 6,577 with 624 stars.

README

ChatPassport

Switch assistants. Keep your context.

Continue a conversation in ChatGPT, Claude, Gemini, or DeepSeek without rebuilding its context by hand. ChatPassport reads the conversation you choose, prepares a draft in another assistant, and leaves the final review and Send click to you.

Get started · Capture modes · What transfers · Privacy · Build from source

The screenshots use demonstration conversations. ChatPassport works alongside the supported websites; it does not replace their accounts or interfaces.

Why use it?

When a chat gets long, you hit a limit, or you want a second assistant's perspective, copying the conversation message by message is tedious. ChatPassport carries the text and code it can read into a new draft, then lets you add your next question before anything is sent.

  • Choose your destination first. Move between ChatGPT, Claude, Gemini, and DeepSeek.
  • Choose your capture depth. Use Quick for currently loaded messages or Load history to scroll for older messages.
  • See what was captured. Review the message count, coverage notice, and estimated draft fit before continuing.
  • Keep control of the draft. If older messages must be left out, ChatPassport asks before inserting anything. You review and send the result yourself.
  • Keep a portable copy. Export Markdown or .chatpassport.json, and import saved JSON later.

ChatPassport runs without a ChatPassport account, API key, developer-operated server, analytics, or tracking.

Get started

The steps below describe ChatPassport 0.1.6. If your installed extension does not show Quick and Load history, check its version in chrome://extensions; Web Store updates can appear after review.

  1. Install ChatPassport from the Chrome Web Store. Open a conversation on a supported website and open the ChatPassport side panel. If the website was already open during installation, refresh that tab once.
  2. Choose another assistant and a capture mode, then click Prepare for [destination]. Check the captured messages and coverage notice. Click Continue in [destination] to open the destination tab.
  3. Type your next question in the destination's normal editor. On the ChatPassport card, click Add context to draft. Confirm any proposed omissions, review the resulting draft, and use the website's own Send button.
Write your next question Review and send

ChatPassport never clicks Send for you. Once context is inserted into the destination editor, that website can read or sync the draft even before you send it.

Choose how much to read

Mode What it does When to use it
Quick Reads messages currently rendered on the page. It does not scroll or wait for history, and marks the result as partial. You need the visible recent exchange quickly.
Load history (default) Scrolls to collect older rendered messages, then restores the page position. It can take up to two minutes. You want more of a long conversation.

While Load history is running, Stop and use collected messages keeps what has been found so far. Both modes report capture coverage. Websites may hide, virtualize, collapse, or load history in ways the extension cannot fully observe; a successful scan does not guarantee every past message was captured. Loading history may also cause the source website's normal network requests.

After capture, More options lets you transfer all captured messages or only the latest 20, 50, or 100. It also contains JSON/Markdown export and JSON import. Exports contain the captured preview, regardless of the range selected for transfer.

What transfers

Included Not included
Conversation text as Markdown (headings, lists, links, tables, fenced code blocks with their language, TeX math), message order and roles, and [image: description] or [image] placeholders for non-decorative images Image and file contents, citation cards, artifacts, hidden reasoning, or messages that the page did not expose

The destination draft has a 48,000-character budget, including your next question. ChatPassport estimates how many complete recent messages fit before you continue, then recalculates using the question you actually typed. If it must omit older messages, it shows the counts and asks you to confirm before changing the editor. It does not silently cut a message in half or summarize it with AI. A destination model may have a smaller context limit.

If the website rejects an insertion, ChatPassport shows the original question and offers Copy original question for recovery. Copy saved context in the side panel provides a manual fallback for a pending transfer within its copy-size limit.

Save or import a conversation

In More options, choose Save JSON for a portable .chatpassport.json file or Save Markdown for a readable .chatpassport.md file. To reuse a saved conversation, choose Import saved conversation (.chatpassport.json) and then a destination.

Files are processed locally. JSON capture, export, and import have a 25 MiB limit; oversized captures retain only complete messages that fit and show a partial-coverage warning. See the JSON format for the schema and compatibility notes.

Privacy and control

ChatPassport reads a conversation only after you choose Prepare for [destination]. It requests access only to the supported assistant domains and uses Chrome's scripting, storage, sidePanel, and alarms permissions for capture, a temporary transfer, the interface, and expiry cleanup.

One pending transfer is stored in chrome.storage.session, bound to the destination tab, and becomes inaccessible after one hour. Expiry is checked on access; cleanup can be delayed while Chrome is asleep or stopped. Exported files remain wherever your browser saves them until you delete them.

No conversation or draft is sent to a ChatPassport server. The source and destination websites handle their own pages and requests under their own policies. Read the full privacy policy.

Supported websites

Assistant Website
ChatGPT chatgpt.com, chat.openai.com
Claude claude.ai
Gemini gemini.google.com
DeepSeek chat.deepseek.com

ChatPassport needs a browser that supports Chrome Manifest V3 and the side panel. You still need access to the assistant websites themselves. Their page layouts can change; if capture or insertion stops working, report the affected website and behavior without including private conversation text.

Build from source

Requires Node.js 20+ and pnpm.

pnpm install
pnpm check
pnpm release:chrome

pnpm check runs type checking, tests, and a production build. pnpm release:chrome also creates the upload-ready ZIP in release/. For a local browser install, open chrome://extensions, enable Developer mode, choose Load unpacked, and select .output/chrome-mv3 after building. The repository root is source code, not an unpacked extension.

Contributing

Issues and pull requests are welcome. For a compatibility report, include the assistant, the action you took, the result you saw, and any Chrome extension errors, with personal conversation content removed. Run pnpm check before opening a pull request.

License

ChatPassport is released under the MIT License.

Star history

If ChatPassport is useful, give the project a star. The chart below shows new stars per UTC day over the past 60 days.

⭐ Star ChatPassport · View stargazers

Nearby in the ranking
Similar items