Commandcode

Command Code API provider for opencode. Use Claude, GPT, Gemini, DeepSeek, Qwen, Kimi, GLM, MiniMax, Step, and other models through a single API key.

This package keeps a bundled model catalog current via CI. You do not need a local command-code CLI. Catalog patches publish automatically after a green PR merges to main.

Previously published as @brainervirus/opencode-commandcode (and before that @brainervirus/commandcode-go-opencode-provider). This fork publishes as @telsoncc/opencode-commandcode — use this name instead.

Credits

This package is based on FanFan4204/opencode-commandcode-provider. That work started from brent-weatherall/opencode-commandcode-provider by Brent Weatherall. Thank you both — FanFan for the OpenCode provider this repo continues, and Brent for the original plugin, catalog extraction.

Forked and currently maintained by telsoncc (OpenCode 2.x support and ongoing catalog updates).

What this package adds

• Bundled models.json is the default runtime catalog (no local CLI scrape).
• CLI cost extraction can fail (as on command-code@1.38.x) without dropping models.
• Official docs fill missing costs; remaining paid gaps use models.dev as a reference. Command Code free SKUs stay $0.
• Vision vs text-only comes from the Command Code CLI catalog (inputModalities on every SKU). models.dev only adds extra inputs (video/audio/pdf) when it matches.
• Reasoning effort variants on models that declare reasoningEfforts.
• Quiet OpenCode startup (diagnostics go to startup.json, not stdout).
• Live catalog refresh on the 2.0 entrypoint: retired models are pruned and new ones are published with models.dev estimates (marked (est.)) until the next catalog sync replaces them.
• Native OpenCode 2.x entrypoint (plugin/opencode2, also served as server): in-memory provider inventory via ctx.provider.transform (no config writes) plus a commandcode integration for /connect. See docs/opencode-2.md.

Quick Start

Use one entrypoint per host. OpenCode 1.x loads the classic plugin; OpenCode 2.x loads the 2.0 entrypoint from the same package.

OpenCode 1.x — install the plugin

{
  "plugin": ["@telsoncc/opencode-commandcode@latest"]
}

Pin a version instead of @latest if you do not want automatic catalog patches.

file:// checkouts are not updated by npm; git pull after CI commits, or switch to the npm plugin line.

OpenCode 1.x — provider transport (Command Code Provider API)

This plugin supplies model metadata. Point OpenCode at Command Code's documented Provider API:

{
  "plugin": ["@telsoncc/opencode-commandcode@latest"],
  "provider": {
    "commandcode": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Command Code GOAT",
      "env": ["COMMANDCODE_API_KEY"],
      "options": {
        "baseURL": "https://api.commandcode.ai/provider/v1"
      }
    }
  }
}

OpenCode 2.x — install the plugin (no provider block needed)

{
  "plugins": ["@telsoncc/opencode-commandcode"],
}

The host resolves the package's ./server export, which serves the 2.0 entrypoint. (Do not append /plugin/opencode2 in config: subpaths are treated as local paths, not npm specifiers.)

The 2.0 entrypoint registers the commandcode provider (native openai-compatible runtime, Provider API base URL) and its 71-model catalog in memory — do not keep a hand-written provider/providers.commandcode block (including models dumped by bun run sync -- --update-global): it fights the plugin-owned inventory. sync --update-global is 1.x-only. Full guide plus local-clone loading: docs/opencode-2.md.

Pin a version instead ("@telsoncc/opencode-commandcode@<version>") if you do not want automatic catalog patches.

Connect

Run /connect in opencode, search for Command Code, and enter your API key, or set COMMANDCODE_API_KEY.

Select a model

/models

On 2.x the catalog carries no release dates, so filter the picker by provider Command Code if the global list looks empty.

Live catalog refresh (2.x)

The 2.0 entrypoint re-checks the Provider API (/models, no auth) on setup and every 6h on a process-shared timer:

• ids the API retired are pruned from the inventory;
• ids missing locally are published immediately with models.dev cost/capability estimates, named "<model> (est.)", and replaced by exact data on the next catalog sync;
• ids with no estimable data stay reported (not published) until the sync.

Estimates are list prices, close to but not equal to Command Code billing. startup.json records catalogSource: "remote", pendingNewCount, and estimatedCount. Opt out with disableModelSync: true in ~/.config/opencode/opencode-commandcode.json.

Optional local CLI override

Maintainers only. OpenCode will scrape a local command-code install when COMMANDCODE_PACKAGE_PATH or commandCodePackagePath in ~/.config/opencode/opencode-commandcode.json is set.

Development

git clone https://github.com/telsoncc/opencode-commandcode.git
cd opencode-commandcode
bun install
bun run check            # oxlint + oxfmt + bun test + tsc (same stack as workit)

bun run sync -- --remote  # refresh models.json + manifest.json from command-code@latest

CI (.github/workflows/catalog-sync.yml) opens a PR every 6 hours when Command Code ships a new catalog. That PR, and the post-release chore/manifest-sync-v* PR, auto-merge after check (test), check (typecheck), check (lint), check (format), and check (pack) are green. .github/workflows/release.yml then runs semantic-release (npm publish + GitHub Release + tag). Do not push to main.

The GitHub Actions secret name is NPMJS (same as workit). It is mapped to both NPM_TOKEN and NODE_AUTH_TOKEN. Use an npm Automation token (bypasses 2FA). A login token from ~/.npmrc fails CI with EOTP. Catalog PRs get a real CI run when RELEASE_SYNC_TOKEN (or CATALOG_PUSH_TOKEN) is a PAT; GITHUB_TOKEN can open the PR but GitHub will not start workflows from that event.

License

MIT — see LICENSE. Original copyright Brent Weatherall; modifications copyright telsoncc.