jev-chat-jarvis: Jev Chat Assistant – Android/Windows/macOS Chat Analysis Tool
Jev Chat Assistant is a cross‑platform app that reads chat screens via accessibility services, uses a judgment model to assess intent and danger, and offers draft replies without auto‑sending, with configurable model endpoints. 7,255 GitHub stars in 11 days since jev-chat/jev-chat-jarvis was created (639.9/day). 1 so
42Fact cited ·Analysis our interpretation ·Estimate forward-looking · [n] links to the source
4 claims drafted by openai/gpt-oss-20b were kept because each quote was found verbatim in its source; 2 unverifiable claims were discarded. Statements marked AI await human review.
What happened
- Fact Jev reads chat content only via system accessibility services and does not modify or hook the chat application.[1]AI
不 hook、不改包、不走任何 App 的接口或账号、不读数据库,只用系统无障碍服务读「屏幕上正在显示的对话」
- Fact The judgment model outputs the other party’s true intent, danger level (1–9), what they want, whether to reply immediately, and the best action, in about one second.[1]AI
判断模型一次给出:对方真实意图、危险等级(1–9)、对方要什么、该不该马上回、最佳动作。约 1 秒,带把握度。
- Fact Reply drafts are generated by a separate model and sorted by the judgment model, with three conversational suggestions provided.[1]AI
生成模型起草 3 条口语化候选,判断模型按「最合适」排序并给出占比。
- Fact The GitHub repository jev-chat/jev-chat-jarvis was created on 21 Sept 2026 and had 7,255 stars on 2 Oct 2026.[1]
- Fact Its description reads: “装在手机上的对话副驾:在 QQ / X / 飞书里读懂对方、给出候选回复、一键填入输入框,发不发由你。非侵入,只读屏幕,不 hook 不改包。”[1]
- Fact That is 7,255 stars in 11 days, an average of 639.9 per day.[1]
Why it matters
- Analysis Integrate Jev’s judgment and reply APIs into their own applications, customize model endpoints, and extend adapters for additional chat platforms.AI
- Analysis Automatically detected from public activity data. Not yet reviewed by an analyst.
Should I care?
What changed?
- Fact Jev reads chat content only via system accessibility services and does not modify or hook the chat application.[1]AI
不 hook、不改包、不走任何 App 的接口或账号、不读数据库,只用系统无障碍服务读「屏幕上正在显示的对话」
- Fact The judgment model outputs the other party’s true intent, danger level (1–9), what they want, whether to reply immediately, and the best action, in about one second.[1]AI
判断模型一次给出:对方真实意图、危险等级(1–9)、对方要什么、该不该马上回、最佳动作。约 1 秒,带把握度。
Who is affected?
- Analysis Developers building chat‑assistant or moderation tools on Android, Windows, or macOS who need a privacy‑preserving, non‑invasive solution for analyzing chat content and generating reply suggestions.AI
What can developers do with it?
- Analysis Integrate Jev’s judgment and reply APIs into their own applications, customize model endpoints, and extend adapters for additional chat platforms.AI
Is it production-ready?
Insufficient evidence
Analysis Not assessed.
What would make it important?
Not yet assessed.
What could prevent adoption?
Not yet assessed.
Technical change
- Fact Jev reads chat content only via system accessibility services and does not modify or hook the chat application.[1]AI
不 hook、不改包、不走任何 App 的接口或账号、不读数据库,只用系统无障碍服务读「屏幕上正在显示的对话」
- Fact The judgment model outputs the other party’s true intent, danger level (1–9), what they want, whether to reply immediately, and the best action, in about one second.[1]AI
判断模型一次给出:对方真实意图、危险等级(1–9)、对方要什么、该不该马上回、最佳动作。约 1 秒,带把握度。
- Fact Reply drafts are generated by a separate model and sorted by the judgment model, with three conversational suggestions provided.[1]AI
生成模型起草 3 条口语化候选,判断模型按「最合适」排序并给出占比。
Architecture
No architecture has been documented for this signal.
Developer impact
Developer implications have not been assessed yet.
Difficulty has not been estimated yet.
Adoption & momentum
No adoption indicators recorded.
Timeline
First seen by the Astrava pipeline[1]
Limitations & risks
- Fact The app currently does not support WeChat Android version, which has been fully removed from collection.[1]AI
微信 Android 版已全面下架,不再采集或处理微信内容。
- Analysis Generated from activity metrics only; verify the linked sources before acting on it.
Competing approaches
No competing approaches documented yet.
Who is building on it
No independent projects or organisations are documented as building on this yet.
Evidence & sources
Primary source (1)
Published by the people who built it: release notes, repositories, papers, official docs.
- [1]jev-chat/jev-chat-jarvis repository(opens in a new tab)
装在手机上的对话副驾:在 QQ / X / 飞书里读懂对方、给出候选回复、一键填入输入框,发不发由你。非侵入,只读屏幕,不 hook 不改包。
Secondary source (0)
Independent reporting, third-party benchmarks, integrations by other projects.
None recorded.
Community signal (0)
Discussion threads and social posts. Evidence of attention, not of correctness.
None recorded.
How the score of 42 was calculated
| Factor | Share of maximum | Points |
|---|---|---|
| NoveltyRubric level 2/4 (Moderate). Heuristic (automated): first published 11 days ago. | 7.5/15 | |
| Technical significanceRubric level 3/4 (High). AI-assisted (openai/gpt-oss-20b), pending human review: Provides a modular, privacy‑focused architecture that separates judgment, reply, and vision, enabling developers to plug in different LLMs and adapt to multiple chat platforms. Evidence: “判断 / 回复 / 视觉三路分别可填。” | 11.3/15 | |
| Developer relevanceRubric level 1/4 (Low). Heuristic (automated): no installable package detected; topic is not primarily developer-facing. | 3.8/15 | |
| Adoption velocityIndependent integrations: unknown (0 pts). GitHub stars: 639.9 per day since creation; full marks at 1000/day, log-scaled (4.7/5). | 4.7/10 | |
| Community momentumFewer than three periods of activity data, or no baseline activity to compare against. | 0/10 | |
| Source credibilityPrimary source present; mean source-tier weight 1.00 (primary 1.0, secondary 0.7, community 0.35). | 10/10 | |
| Cross-source confirmationNo independent publisher has confirmed this yet. | 0/10 | |
| Evidence strengthMet 2/4: primary source ✓, runnable artifact (code, weights or package) ✓, reproducible or independent benchmark ?, technical documentation or paper ✗. 1 unknown (scored 0). | 5/10 | |
| Time to impactNo time-to-impact estimate yet. | 0/5 | |
| Hype penaltyNo hype indicators detected. | 0 | |
| Breakout Score | 42 |
Low confidence (0.48) — 60% of measured factors have data. 0 independent non-community publishers. At least one primary source. Evidence strength 5/10. Capped at Low: no independent, non-community source yet.
An analytical ranking derived from observable signals using model breakout-v1.2. It is not a prediction and not a measure of quality. How it works