SuperCoder AI Agent

SuperCoder

A local-first, open-source coding agent for your desktop. Bring your own LLM key; your code stays on your machine and only ever leaves to the model provider you choose — no middleman service, no lock-in.

Turn on the optional Context Engine and the agent navigates large codebases structurally — tree-sitter → vector + call-graph + BM25 retrieval — instead of guessing.

SuperCoder has been reimagined from the ground up. The original (2024) autonomous-dev pipeline is frozen under v1/ — preserved, not maintained or built.

---

Why SuperCoder

• Local-first & fully open. A desktop app, not a cloud product. Your source never transits a vendor backend — requests go straight from your machine to the provider whose key you configured.
• Bring your own model. The agent speaks the OpenAI chat-completions and Anthropic Messages APIs natively — no translation proxy.
• Graph-aware code understanding (optional). The Context Engine indexes your repo into vector + call-graph + lexical search so the agent can locate code by structure, not just text similarity.
• A real harness underneath. The core is a pure-Rust agent crate with Ask / Plan / Coding modes, subagents, skills, tool approval, and prompt caching. The desktop app is one adapter over it — see ARCHITECTURE.md.

Two ways to run

SuperCoder works the moment you add an LLM key — in-place edits, Ask / Plan / Coding modes, checkpoints and rewind, diff review, an interactive terminal, and a file explorer. Zero backend required.

Flip on the Context Engine (Settings → Context engine) for graph-aware, repo-scale retrieval. It runs locally via docker compose and the agent's codebase_search / codebase_graph tools query it. See services/context-engine/README.md.

Getting started

Prebuilt downloadable binaries are coming. For now, build from source.

Prerequisites

• Rust (stable) and the Tauri 2 system prerequisites for your OS (WebView / build tooling).
• Node.js 20+ and npm.
• (Optional, for the Context Engine) Docker with Compose.

Run the app

cd apps/desktop
npm install
npm run tauri:dev      # development
# or
npm run tauri:build    # produce a release bundle

On first launch, open Settings and add an LLM provider (base_url + api_key + model). Then create a session, pick a folder and a mode, and go.

(Optional) Run the Context Engine

cd services/context-engine
cp .env.example .env    # set SUPERCODER_OPENAI_API_KEY (server-side embedding key)
docker compose up -d --build

Then enable Settings → Context engine in the app. Full instructions: services/context-engine/README.md.

Repository layout

crates/
  agent/             Rust agent core — the harness (loop, tools, modes, subagents)
  git-ops/           Checkpoint / diff / restore over the working tree
apps/
  desktop/           Tauri 2 + React desktop app (thin adapter over the core)
services/
  context-engine/    Optional Go indexing service (tree-sitter → Qdrant + FalkorDB + BM25)
v1/                  Legacy 2024 codegen pipeline — frozen, not built

See ARCHITECTURE.md for how these fit together.

Roadmap

Present-tense — what works today — is above. Next:

• Prebuilt releases & installers (the CI to produce them lands next).
• Benchmarking the harness. A headless runner over the same agent core, with reproducible per-task execution sandboxes, to measure the harness as an equalizer across models and to validate the graph-retrieval localization claim.
• Broader provider support (the provider abstraction is built to grow).

Contributing

Contributions are welcome — see CONTRIBUTING.md for dev setup and repo conventions, and CODE_OF_CONDUCT.md.

• Bugs & features: GitHub Issues.
• Questions & ideas: GitHub Discussions.
• Security: please report privately — see SECURITY.md.

License

MIT © TransformerOptimus.