Codeep — Autonomous AI Coding Agent<br>GitHubGet Started
v2.1.4<br>Go Deep<br>into Code.<br>CODEEPCODEEPCODEEP<br>An autonomous AI agent that reads your whole project, runs commands, and writes & verifies code on its own — going deeper than autocomplete. Bring any model, hosted or local.<br>Get StartedView on GitHubVS Code Extension<br>Works with any model<br>ClaudeGeminiDeepSeekGLMOpenAIOllamaOpenRouter
50+Built-in skills<br>9+AI providers<br>30+Slash commands<br>FreeOpen source
Latest on npm<br>v2.1.4<br>npm install -g codeep
Downloads / month<br>4.3K<br>last 30 days
GitHub stars<br>17<br>open source · Apache 2.0
Latest release<br>Today<br>May 22, 2026
How it works<br>From prompt to verified code<br>Codeep handles the full development loop autonomously — no context switching, no copy-pasting.
01
You describe the task<br>Type your request in plain language. No special syntax required. Codeep reads your project structure and key files automatically to understand the context.
input<br>~ $ > Refactor the auth module to use JWT tokens
02
Agent plans and executes<br>Codeep breaks the task into steps, executes them one by one — editing files, running commands, fetching documentation. You watch it work in real-time.
terminal<br>[1/4] Reading src/auth/<br>[2/4] Creating jwt.ts<br>[3/4] Updating middleware<br>[4/4] Running typecheck
03
Verified and ready<br>After making changes, Codeep runs your build, tests, or typecheck to verify everything works. If something breaks, it fixes it automatically.
output<br>✓ No type errors<br>✓ All tests pass<br>2 files created, 1 modified
Use cases<br>One command for every task<br>Codeep understands your codebase and executes real work — not just suggestions.
Fix a bugAdd a featureRefactor codeWrite testsReview changes
terminal<br>~$codeep fix the login form - validation isn't triggering on submit<br>▸Scanning project — Reading src/components/LoginForm.tsx
▸Identifying issue — Missing onSubmit handler — e.preventDefault() not called
▸Applying fix — Patching LoginForm.tsx (2 lines)
▸Verifying — ✓ Build passed · ✓ No type errors
✓ Fixed. The form handler was missing e.preventDefault(), causing the page to reload before validation ran.
Capabilities<br>Everything you need to code with AI<br>From autonomous agent tasks to git workflows — Codeep handles the full development loop from your terminal. Green “New” tags mark what shipped in 2.0.
Autonomous Agent<br>Reads files, runs shell commands, edits code, and auto-verifies the result (build / typecheck / tests) — all in one loop. You describe the goal; Codeep handles the steps.
File editsShell execAuto-verify
New<br>Plan Mode<br>Preview the agent's full plan — files it'll touch, commands it'll run, risk it self-assesses — before any code changes. `/plan ` → review → `/go` to execute.
/plan/goPre-execution review
Full Project Context<br>Auto-scans your project structure, detects language and framework, reads key files. No manual context setup. Project memory + lifecycle hooks persist across sessions.
Auto-scan/memory.codeep/hooks/
New<br>Personalities<br>Switch agent tone mid-conversation: concise, verbose, security-paranoid, senior-reviewer, junior-mentor, ship-it. Or drop your own Markdown file in `.codeep/personalities/`.
6 presetsCustom .mdPer-project
Git Workflows<br>AI review of your diff — bugs, security, performance, edge cases. Semantic commit messages. Works with any git repository. Hooks let you gate commits on custom checks.
/review/diff/commit
New<br>OpenRouter<br>100+ models from Anthropic, OpenAI, Google, Meta, DeepSeek, Mistral via one key. Authoritative cost from `usage.cost` — your dashboard sees the same numbers as your invoice.
100+ modelsReal per-call cost/openrouter
New<br>Prompt Caching<br>Automatic on Anthropic (and OpenRouter → Anthropic routes). Cache hits bill at 0.1× input rate; typical agent session is 60-90% cheaper. No config, just upgrade.
Anthropic60-90% cheaper/cost shows savings
New<br>MCP Marketplace<br>Curated catalog of 12 Model Context Protocol servers — filesystem, GitHub, Postgres, browser, Slack, more. One command install. Sampling guards keep credits safe.
12 serversstdio + HTTP/mcp install
New<br>Skill Marketplace<br>Write a SKILL.md (Markdown + YAML frontmatter), agent auto-invokes when triggers match. Publish to share — `/skills install /` from any machine. Claude Code-compatible.
SKILL.mdBrowse + publishAuto-invoke
New<br>VS Code & Zed (ACP)<br>Native agent inside VS Code and Zed via Agent Client Protocol. Same CLI binary, same commands, native diff viewer, image paste, manual confirmation mode.
VS CodeZedACP
New<br>Personal Dashboard<br>Link with `codeep account` to sync sessions, profiles, skills, and API keys across machines. Track cost-per-model, file bug reports, manage devices — all at codeep.dev.
Multi-device syncCost trackingBug reports
Secure & Local<br>API keys stored locally (encrypted at rest, synced encrypted if you opt in). Dangerous commands require confirmation. Full undo history. `/checkpoint` snapshots for safety.
Local keys/checkpoint/undo
MCP marketplace<br>One command, real integrations<br>Codeep...