Skip to content

arpjw/nimbus

BranchesTags

Repository files navigation

Nimbus

Autonomous software engineering, stratified.

PyPI version Python 3.11+ MIT License

Nimbus plans, implements, and reviews code against real repositories — entirely on its own. From task description to merged PR.

get-nimbus.com · docs.get-nimbus.com · api.get-nimbus.com

Status

Feature State
Core pipeline (plan, implement, verify, PR) Stable
CLI (nimbus run, nimbus diff, nimbus chat, nimbus search) Stable
Web IDE (Monaco, terminal, file tree) Beta
Slack / Linear integrations Beta
Skill marketplace Beta
Agent pipelines Beta
GitHub Actions (arpjw/nimbus-action@v1) Experimental
Python SDK (nimbus-sdk) Experimental
TypeScript SDK (@nimbus-ai/client) Experimental
Chrome Extension Experimental
Homebrew tap Experimental
VS Code Extension Experimental

Install

pip install nimbus-ai

Set your API keys:

export ANTHROPIC_API_KEY=sk-ant-...
export VOYAGE_API_KEY=pa-...

Then run in any git repo:

cd your-project
nimbus

What Nimbus does

Nimbus runs a 10-phase pipeline on every task:

Clone -> Index -> Plan -> [Approve] -> Implement -> Verify -> [Diff] -> Review -> PR -> Memory

Every task goes from a plain English description to a merged pull request. Nimbus indexes your codebase with hybrid RAG (voyage-code-2 + BM25), plans with Claude Opus, implements with Claude Sonnet, verifies against your real test suite, and opens a PR with a self-review attached.

Terminal commands

nimbus                          # Interactive REPL — start here
nimbus run "task description"   # Run a task directly
nimbus run --tdd "task"         # TDD mode: generate failing tests first, implement to pass
nimbus run --agent <name>       # Run a built-in agent
nimbus chat                     # Ask questions about your codebase
nimbus diff [rev-range]         # Review any diff (or pipe from stdin)
nimbus diff --staged            # Review staged changes before committing
nimbus search "query"           # Semantic search over your codebase
nimbus install-hooks            # Install pre-commit review hook
nimbus uninstall-hooks          # Remove pre-commit hook
nimbus replay                   # Replay a past session
nimbus watch                    # Ambient mode — suggestions as you code
nimbus pair                     # Per-file suggestions on every save
nimbus memory                   # View per-repo codebase memory
nimbus explain <file>           # Explain a file in plain English
nimbus agents                   # List built-in agents
nimbus skills list              # List available skills
nimbus skills search <query>    # Search the skill marketplace
nimbus skills install <name>    # Install a community skill
nimbus skills publish           # Publish a skill to the marketplace
nimbus pipeline list            # List configured pipelines
nimbus pipeline run <name>      # Run a pipeline manually
nimbus health                   # Codebase health scan — 6 metrics with scores
nimbus models                   # Show configured models for each role
nimbus plugin list              # List installed CLI plugins
nimbus plugin install <pkg>     # Install a plugin from PyPI
nimbus plugin uninstall <pkg>   # Remove a plugin
nimbus plugin run <name> <cmd>  # Run a plugin command
nimbus reindex <repo-id>        # Force re-index a repository
nimbus --version                # Show version

Built-in agents

Run any agent with nimbus run --agent <name>:

Category Agents
Security security-audit secret-scanner dependency-cve
Quality type-safety error-handling dead-code-eliminator complexity-reducer naming-consistency
Testing test-coverage integration-test-builder mutation-tester
Docs api-documenter readme-architect inline-documenter
Performance query-optimizer async-converter
Infrastructure ci-builder docker-hardener
Architecture feature-flags observability-agent

Custom models

Configure which model to use for each role in ~/.nimbus/config.toml:

[models]
planner     = "claude-opus-4-6"
implementer = "claude-sonnet-4-6"
reviewer    = "claude-sonnet-4-6"

Supported providers: Anthropic (default), OpenAI (set OPENAI_API_KEY — reviewer uses gpt-4o automatically).

Bring your own key (BYOK)

Store your own Anthropic or Voyage key on your account so Nimbus uses it for all tasks instead of billing against your plan:

curl -X POST https://api.get-nimbus.com/keys/byok \
  -H "X-API-Key: nk_..." \
  -H "Content-Type: application/json" \
  -d '{"anthropic_key": "sk-ant-...", "voyage_key": "pa-..."}'

Keys are stored encrypted at rest. When a BYOK key is configured the monthly spend cap does not apply.

Plugins

Extend the Nimbus CLI with community plugins:

nimbus plugin install nimbus-plugin-jira
nimbus plugin list
nimbus plugin run jira <cmd>

Plugins are Python packages prefixed with nimbus-plugin- on PyPI.

Web IDE

Open any GitHub repo in the browser — no install required:

get-nimbus.com/ide

Monaco editor, live terminal, file tree, and Nimbus chat panel. Runs on isolated containers per session.

Integrations

Surface How
GitHub PRs/Issues Comment /nimbus on any issue or PR
Slack /nimbus run "task" slash command
Linear Assign issue with nimbus label to auto-implement

Per-repo configuration

Drop a .nimbus.toml in your repo root to override defaults:

[paths]
denied = ["secrets/", "*.pem"]

[branches]
protected = ["main", "release/*"]

[models]
planner     = "claude-opus-4-7"
implementer = "claude-sonnet-4-6"

[policy]
require_human_approval = true
block_on_severity      = "error"

Configuration

~/.nimbus/config.toml:

[core]
default_branch = "main"
editor = "nvim"

[hooks]
block_on = "high"               # pre-commit hook severity threshold

[models]
planner = "claude-opus-4-6"
implementer = "claude-sonnet-4-6"
reviewer = "claude-sonnet-4-6"

Self-hosting

git clone https://github.com/arpjw/nimbus
cd nimbus

# Backend
cd backend
pip install -r requirements.txt
cp .env.example .env   # add your API keys
uvicorn main:app --reload

# Frontend
cd ../frontend
npm install
npm run dev

Ecosystem

Package Install Status
CLI pip install nimbus-ai Stable
Web IDE get-nimbus.com/ide Beta
Python SDK pip install nimbus-sdk Experimental
TypeScript SDK npm install @nimbus-ai/client Experimental
GitHub Action uses: arpjw/nimbus-action@v1 Experimental
Chrome Extension arpjw/nimbus-chrome Experimental
VS Code Extension arpjw/nimbus-vscode Experimental

Known limitations

Redis + ARQ task queue (not yet implemented): Tasks currently execute in-process on the FastAPI request handler. Under concurrent load this blocks the event loop. A proper ARQ worker queue with Redis pub/sub is planned but not yet shipped. Workaround: run a single uvicorn worker and limit concurrent tasks via the rate limiter.

Task reconciliation on restart (not yet implemented): If the server restarts while a task is in-flight, the task stays in its current phase forever. There is no reconciler that resets stuck tasks to FAILED on startup. Workaround: manually PATCH the task status, or set a short timeout in your client and retry.

Changelog

v1.5.0 — May 2026

  • BYOK (bring your own key) — store encrypted Anthropic/Voyage keys per account; spend cap bypassed when BYOK is active
  • Incremental repository indexing — skip unchanged files via SHA256 content tracking
  • Per-repo .nimbus.toml config — override models, deny paths, protect branches
  • Editable plans — PATCH /tasks/{id}/plan lets users modify file changes before approval
  • Sandboxed verification — SANDBOX_VERIFICATION=true runs ruff/pytest/mypy inside Docker
  • Cross-provider self-review — reviewer uses gpt-4o when OPENAI_API_KEY is set
  • Static analysis override — APPROVE verdict upgraded to REQUEST_CHANGES when ruff reports errors
  • Skill safety — prompt injection scanning and Haiku moderation on skill publish
  • Structured JSON logging via structlog; optional Sentry integration

v1.4.0 — Apr 28, 2026

  • Web IDE at get-nimbus.com/ide — Monaco editor, xterm.js terminal, isolated containers per session
  • Nimbus chat panel in the IDE

v1.3.0 — Apr 27, 2026

  • nimbus run --tdd — TDD mode: generate failing tests first, implement to pass
  • nimbus health — codebase health scan with 6 metrics
  • nimbus models — show configured models per role
  • CLI plugin system — nimbus plugin install/list/run, entry-point based discovery
  • TaskMetrics model tracking duration, success rate, cost per task

v1.2.0 — Apr 26, 2026

  • nimbus chat — conversational codebase Q&A with retrieval-backed citations
  • nimbus diff — review any diff; pipe from stdin or pass revision range; --exit-code for CI
  • nimbus install-hooks — git pre-commit hook that runs nimbus diff --staged
  • nimbus search — semantic search over indexed codebase
  • Skill marketplace — nimbus skills search/install/publish
  • Agent composition pipelines

v1.1.0 — Apr 26, 2026

  • Interactive terminal mode (nimbus) with live diffs, session replay
  • 20 built-in agents across 7 categories
  • nimbus watch ambient mode, nimbus pair per-file suggestions
  • Slack slash commands, Linear webhook integration
  • GitHub OAuth — personal dashboard, account-tied API keys

v1.0.0 — Apr 25, 2026

  • Core 10-phase pipeline
  • GitHub App with /nimbus comments on issues
  • Self-improving PR reviewer
  • Persistent per-repo memory
  • Parallel implementation workers

Links

Built by Arya Somu · Monolith Systematic LLC · MIT License

About

Your codebase, on autopilot.

get-nimbus.com

Topics

claude rag autonomus swe-agent

Resources

Readme

License

MIT license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases 4

Nimbus v1.4.0 Latest
Apr 28, 2026
+ 3 releases

Packages

 
 
 

Contributors

Languages