Flick
Local stdio MCP server that executes whole browser and macOS goals with a Jev-driven observe-decide-act loop.
Overview
Flick is a local stdio MCP server for fast computer use. The calling agent supplies a whole goal, exact input values and an observable completion condition, and Flick runs the observe, decide, act loop with Jev, returning the result, rather than the agent spending an LLM turn on every click. It has a TypeScript core, a Playwright browser driver and a Swift accessibility driver for native macOS apps.
Installation is designed to be done by your coding agent from a prompt and an INSTALL.md setup contract, or manually with Node 22+ and a TypeSafe key: a setup script, an MCP config generator, a doctor command and a live smoke test against a disposable local form. No Flick account or dashboard is needed. The README recommends delegating whole goals with explicit inputs and completion conditions.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which action on which observed element should be taken next toward the goal?(inferred) | — |
| Yes / no | Is the completion condition satisfied?(inferred) | yes · no |
How it works
- Agent calls Flick with goal, values and completion condition
- Flick observes the browser or macOS app
- Jev decides the next action; Flick executes
- Repeat until the completion condition holds; return the result
Strengths
- Delegates whole goals instead of per-step LLM turns
- Browser and native macOS support
- No account required
Caveats
- Requires a TypeSafe key
- Native macOS support needs a build step
Getting started
node scripts/setup-agent.mjs, set TYPESAFE_API_KEY in .env.local, npm run --silent mcp:config, then add the flick entry to your MCP client.At a glance
- Type
- MCP server
- Status
- beta
- Language
- TypeScript
- Runtime
- MCP server (Node, Playwright, Swift)
- Has code
- yes
- Has tests
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Local stdio MCP server that executes a whole browser or macOS goal from the agent's goal, values and completion condition.
MrJev/awesome-jev
Frequently asked questions
What does the agent send Flick?
A goal, exact values and an observable completion condition.
Does it support desktop apps?
Yes, native macOS via a Swift accessibility driver.
Do I need a Flick account?
No, only a TypeSafe key.
Related projects
Browser and computer use for coding agents via CLI and MCP, with Jev's calibrated judgments and a code-owned control loop.
MCP serverbetaTypeScriptMCP server and libraries that drive Chrome over CDP and hand a decision model like Jev a bounded menu of page actions.
ProjectproductionTypeScriptRust MCP server for desktop control via accessibility trees, with an optional decision model (Jev or compatible) for small choices.
MCP serverproductionRust- Jev Browser★ 335
Headless browser automation via MCP server, CLI or library: Jev picks one action per step and scores goal-met and stuck probabilities.
MCP serverbetaTypeScript Rust CLI for desktop computer use via accessibility trees, with optional jev-desktop scripts letting Jev pick the control and action.
Agent skillbetaRustmacOS computer-use package: a background app driver for any agent, plus a loop where Jev picks each UI action for a bounded subtask.
ProjectbetaPython