shengjidaguai-china/goutoujunshi-jev-chat
Organization repository on GitHub
狗头军师 Chat：Mac 微信读屏、关系分析与回复草稿悬浮窗
- Forks
- 30
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Tools & CLIs
Goutoujunshi Jev Chat
Cross-platform chat copilot that OCRs a visible conversation, has you correct the transcript and speakers, then uses Jev or DeepSeek to pick one of seven reply strategies before drafting replies.
- Stars
- 376
- Language
- Python
- License
- MIT
- Published by
- Community
- Author
- shengjidaguai-china
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- Python
Highlights
- Strategy judgment is a separate choice: Jev by default, DeepSeek as an independent judge, or automatic with fallback to the reply model.
- Nothing is sent until you confirm: recognized text and speaker are checked first, auto-analysis ships disabled, and the program never presses send for you.
- macOS needs Python 3.12 and uv plus Screen Recording and Accessibility grants; Windows ships an executable ZIP with RapidOCR, DeepSeek, or OpenRouter vision.
- DeepSeek mode weighs the first token across seven strategies over three rotated passes and marks the weights unusable when they come out unstable.
Watch out
Preview builds on all three platforms: macOS is a source ZIP with no signed `.app`, Windows is a ZIP that must be fully extracted, and Android is a debug APK on which WeChat capture is disabled. Real-device acceptance for Windows and for the non-WeChat Android paths is still outstanding, and the percentages shown are relative candidate weights rather than intent probabilities. The repository body is MIT, with third-party notices for the vendored window module and the Android and Windows references. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
No, not in the current build. The Android debug APK explicitly disables WeChat capture, and the README warns against reading the macOS screenshots as Android behaviour. Apps recognised through manual screenshot, plus QQ, X, and Feishu paths, still need real-device verification.
One bounded call: which of seven reply strategies fits the conversation you reviewed. Reply generation is a separate choice routed to DeepSeek or OpenRouter. The candidate percentages are this round's relative recommendation weights and the confidence figure is the model's self-assessment of its own inference, neither of which is the other person's intent probability.
Apple Vision OCR is local by default. Switching image recognition to DeepSeek or OpenRouter sends screenshots of the chat area to that service, and cloud analysis sends the conversation content the same way, so both add provider usage. Reply and vision keys are configured separately and stored in a project-specific macOS keychain entry, and the repository ships no real keys.
On Windows it writes into the current chat draft by window coordinates, checking the session and foreground window but unable to read the input field back to confirm what landed, so the README suggests copying candidates and pasting manually when unsure. The macOS fill-in needs Accessibility permission, and a failed judgment stops reply generation rather than guessing.
Source and preview packages exist for macOS, Windows, and Android, but Android currently cannot capture WeChat chat screens and does not support WeChat. The Windows build targets WeChat 4.x, and the Mac version reads WeChat after you grant screen-recording permission to the terminal that launches it.
Strategy judgment can use TypeSafe Jev, DeepSeek's own judgment, or automatic (Jev first, then the reply model). Replies come from DeepSeek or OpenRouter, and image recognition can use Apple Vision locally by default, DeepSeek, OpenRouter, RapidOCR, or ML Kit. Each key is configured separately and stored in the project's Mac Keychain entry.
Reply style only references current-session lines verified as yours, with no training and no reading of other sessions. The optional relationship profile is saved only after explicit consent and can be viewed, paused, withdrawn, or deleted; it stores limited background, not the whole chat, and no real keys are in the repository.
More Jev resources like this
452GitHub stars
An AGPL MCP server and JSON-first CLI exposing 110 astrology and metaphysics techniques from Horosa to local AI clients, running offline on macOS and Windows with SQLite-backed call history.
Jev Tools & CLIs#community#jevbest
Community
Horace-Maxwell/horosa-skillPython
Horace-Maxwell/horosa-skillPython227GitHub stars
Local-first summarizer for video and posts across 11 or more sources; its optional Jev prefilter scores transcript chunks against include and exclude rules before the summary request is made.
Jev Tools & CLIs#community#jevbest
Community
martinopiaggi/summarizePython
martinopiaggi/summarizePython44GitHub stars
Mac app that turns one short English or Japanese sentence into Ableton Live mixer, transport, clip, note, device, and track edits from a pop-up bar.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
okinaaudio/live-jevPython
okinaaudio/live-jevPython547GitHub stars
Live SEO audit that crawls a site from one homepage URL, checks 52 rules tied to Google Search Central, and asks Jev typed questions about every page to rank fixes.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
AgriciDaniel/jev-seoPython
AgriciDaniel/jev-seoPython27GitHub stars
Local-first trading journal and review harness that is read-only over markets, captures runs into frozen envelopes and evidence packs, and requires explicit human promotion before any rule returns.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
myc0576/SmartMoney-CubPython
myc0576/SmartMoney-CubPython39GitHub stars
Video tool where Jev scores subtitle segments and code assembles a short highlights cut plus a de-watered full cut with timestamped summaries.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
cclank/jevclipPython
cclank/jevclipPythonFrom the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.