Skip to content

configuration.md

Latest commit

 

History

History
92 lines (66 loc) · 4.07 KB

configuration.md

File metadata and controls

92 lines (66 loc) · 4.07 KB
Raw

Configuration

Micracode is configured through environment variables and, in the packaged desktop app, the in-app Settings panel.

  • Desktop app: API keys entered in Settings are persisted to a shared auth file at ~/.micracode/auth.json and applied to the running core.
  • From source: the core backend (@micracode/core) reads a .env at the repo root; the Next.js app picks up NEXT_PUBLIC_* vars at build/dev time.

A working template is committed at .env.example — copy it to .env and edit.

Variable reference

Web (apps/web)

Variable Default Purpose
NEXT_PUBLIC_API_BASE_URL http://localhost:8000 Where the browser calls the core. Must match the port the core binds to.

Core backend (@micracode/core)

Variable Default Purpose
OPENAI_API_KEY — Enables OpenAI models. Set at least one provider key.
GOOGLE_API_KEY — Enables Google Gemini models.
ANTHROPIC_API_KEY — Enables Anthropic models (resolved by the pi coding agent SDK).
PORT 8000 Port the core binds in dev (bun run dev:core). Ignored in the desktop app, which picks a free port.
OPENER_APPS_DIR ~/opener-apps Override for where generated projects are stored. Absolute path.
MICRACODE_CONFIG_DIR ~/.micracode Directory holding auth.json (persisted API keys).
TEMPMD_API_URL https://api.temp.md Override the temporary-preview API endpoint for development or testing. HTTPS is required.

API keys never leave your machine. The browser only learns whether a provider is available (so the model picker can grey out unconfigured ones); it never sees the keys themselves. The pi coding agent SDK also resolves keys from your environment and from ~/.pi/agent/auth.json.

Choosing a provider and model

The chat composer's model picker sends the selected (provider, model) pair with each turn; the selection is remembered in your browser's localStorage, per-browser (not per-project).

GET /v1/models reports the available providers and a minimal built-in catalog (currently OpenAI and Gemini). Because agent sessions accept a free-form model string, the catalog is mostly a convenience for the picker rather than a hard allowlist.

To see the live list — including which providers are currently "available" (a key is configured) — hit:

curl http://127.0.0.1:8000/v1/models

And to check overall health:

curl http://127.0.0.1:8000/v1/health

Changing ports

The defaults assume:

  • Web on :3000
  • Core on 127.0.0.1:8000

To use different ports, change them in two places so they match:

  1. The startup commands — the web port is bun --filter web dev -- -p <port>; the core port is the PORT env var read by dev:core.
  2. The env var NEXT_PUBLIC_API_BASE_URL — point it at the new core URL.

CORS on the core defaults to allowing any origin in dev, so a mismatch shows up as failed requests to /v1/... rather than a browser CORS error — check that NEXT_PUBLIC_API_BASE_URL matches the core's port.

Changing the storage location

By default, projects are written under ~/opener-apps/. To put them elsewhere — for testing, on an external drive, or in a sandbox — set:

OPENER_APPS_DIR=/absolute/path/to/your/folder

The path must be absolute. The core creates the directory if it doesn't exist. See Projects on Disk for what gets written.