Organization repository on GitHub
APUS Reflex Harness — 瞬时执行，深度思考。Model as Harness with hybrid reflex and reasoning models.
- Forks
- 2
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Agents & MCP
APUS Reflex Agent
A developer-preview personal agent harness that turns a sequence of model calls into continuing work: understand the request, choose an action, execute, inspect evidence and recover context. You pick the general model and can optionally connect a Reflex/Jev decision service.
- Stars
- 21
- Language
- TypeScript
- License
- MIT
- Reading time
- 2 min
- Published by
- Community
- Author
- APUS-AI-Lab
- Last verified
- Oct 2026
- Languages
- TypeScript
- Platforms
- macOS
- Capabilities
- Open sourceRequires API key
Highlights
- Model-as-a-harness: APUS turns model calls into work that continues around a goal — understand, choose, execute, inspect evidence and continue.
- You choose the general model for planning and generation; connecting your own Reflex/Jev decision service for structured decisions along the way is optional.
- Developer preview v0.1.3 ships Full and Lite editions; macOS needs 15.0 or later and Mac bundles are ad-hoc signed without Apple notarization.
- The desktop distribution bundles its runtime, so users need no Git, Node.js or separate backend; the Linux Web archive listens on localhost.
- MIT TypeScript repo: 21 stars, 2 forks, created 2026-10-09, pushed 2026-10-17, though the README says the repository is currently private.
Watch out
README reviewed 2026-10-17 and truncated inside a personal-project example, so the full working loop and the Reflex/Jev connection setup are not in the digest. The README itself says the repository is currently private and warns that installation qualification does not establish real-model task completion or recall quality.
Questions about this entry
Short answers drawn from the linked project or page.
A developer-preview personal agent harness from APUS AI Lab: it turns a sequence of model calls into work that continues around a goal, separating model reasoning from execution, permissions and durable session history.
APUS lets you optionally connect your own Reflex/Jev decision service for structured decisions along the way; the general model you choose handles planning and generation, and the README frames the design as model-as-a-harness.
Full bundles the Office engine; Lite keeps the core Agent, Reflex, memory, Codex and PPT workbench while excluding that engine. The README points to an edition-selection document for preview limitations and upgrade behavior.
macOS 15.0 or later for the Mac preview, which uses ad-hoc signatures without Apple Developer ID signing or notarization. The desktop distribution bundles its runtime, so users do not need Git, Node.js or a separate backend.
Yes. The README says to follow the installation and model-configuration guides using your own provider credentials, and it warns that preview builds can break compatibility between versions.
Download from the v0.1.3 release page once you have repository access — the README says the repository is currently private, so downloads require access until an administrator makes it public. GitHub source archives are not installers.
Read SAFETY.md before granting file access. The README notes that installation qualification does not establish real-model task completion or recall quality, and each release reports its tested environments and limitations.
More Jev resources like this
A Jev-first agent loop for DeepSeek Harness: Jev reads the same assembled session and tool history as the main model and either dispatches a saved Action, asks the main model, clarifies, or ends the turn. Requests covered by verified Actions can finish with zero main-model calls.
Jev Agents & MCP#community#github#agent
Community
yinhong-zhou/jevdoTypeScript
yinhong-zhou/jevdoTypeScriptA local stdio MCP server from bgivenb that runs whole browser and macOS goals for your agent: your agent supplies the goal, exact values, and completion conditions, and Flick loops observe, decide, and act with TypeSafe Jev. TypeScript core with Playwright and Swift drivers.
Jev Agents & MCP#community#github#mcp
Community
bgivenb/flick-computer-useTypeScript
bgivenb/flick-computer-useTypeScriptA DeepSeek Harness plugin that exposes Jev's noul, choice and score primitives as agent tools, plus configurable guardrails, retries, response caching, a circuit breaker and Web UI result cards.
Jev Agents & MCP#community#typescript#deepseek
Community
jackie-cqz/dsh-jev-pluginTypeScript
jackie-cqz/dsh-jev-pluginTypeScriptAn OpenCode plugin that prunes stale tool calls and truncates bulky results on the outgoing V2 request using Jev, leaving stored history untouched and failing open when Jev is unreachable.
Jev Agents & MCP#community#typescript#opencode
Hands the agent's browser clicks and typing to Jev: wingman hides Playwright MCP's page-action tools, asks Jev which element fits each step, and returns the step to the agent when it is unsure.
Jev Agents & MCP#community#typescript#mcp
A Claude Code plugin that hands Claude Code's small, frequent decisions to Jev at six points: four event hooks and two routing decisions, each a roughly one-second call that reduces permission prompts, warns when fetched content addresses the agent, re-checks done, and routes effort and model.
Jev Agents & MCP#community#github#claude-code
Community
yinjs/claude-jevTypeScript
yinjs/claude-jevTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.