Pangloss

"All is for the best in this best of all possible worlds" — Voltaire

Pangloss is a multi-model fusion code-generation system. It runs a diverse roster of AI coding agents — different models, and different agentic harnesses — in parallel, has them cross-review each other's work, and synthesizes the best of all possible worlds into one result. It's a bespoke take on OpenRouter's Fusion idea: a loop of diverse, sonnet-level and open-weight models can rival — and in combination exceed — any single frontier model.

How it works

Each run is a loop over four phases. Every agent works in its own git worktree (isolated checkout + branch); agents run on the host, so each tool uses its own existing auth.

feature request
   │
   ▼  PLAN        N models draft independently → a rotating synthesizer merges
   │              them into one canonical plan (+ optional human approval gate)
   ▼  CODE        each model implements the plan in its own worktree, runs the
   │              project's build + tests, iterates to green, commits
   ▼  REVIEW      every model reviews every implementation: score, novel ideas,
   │              gaps, what's still needed
   ▼  SELECT      weighted vote (self-reviews down-weighted) picks a winner and
   │              produces a revision brief (must-fix + ideas to graft + gaps)
   └──▶ if not converged: re-base every agent on the WINNING branch, turn the
        brief into a revision plan, and loop. Stop on convergence or round cap.

The payoff is the cross-pollination: even when several agents pass, the review phase surfaces each one's unique good ideas, and the revision rounds graft them into the kept winner.

The roster — diversity is the product

Agents are <tool> + <model> pairs. The same model through two different harnesses (e.g. Sonnet via claude-code vs. via codex→OpenRouter) genuinely produces different work.

Tool	How it's driven	Examples
claude	Claude Code CLI (claude -p)	claude:sonnet, claude:opus
codex	OpenAI Codex CLI (codex exec)	codex:gpt-5
codex --oss	local open-weight via Ollama	oss:gpt-oss:120b
cursor	Cursor Agent CLI (cursor-agent -p)	cursor:claude-4.5-sonnet, cursor:kimi-k2.5
OpenRouter	any OpenRouter model via codex's custom provider	openrouter:qwen/qwen3-coder, openrouter:z-ai/glm-4.6
gemini	Gemini CLI (gemini -p)	gemini:gemini-2.5-pro (needs GOOGLE_CLOUD_PROJECT)

Drop any of these into a roster ad hoc (--roster "openrouter:qwen/qwen3-coder,claude:sonnet,oss:gpt-oss:120b"), or use a named roster from pangloss.config.json (open-weight-heavy, frontier, sonnet-family, openrouter, …).

Quick start

yarn install
yarn build

# See the roster catalog and named rosters
node dist/cli.js agents

# Check that your roster's CLIs are installed/authed and preview invocations
node dist/cli.js doctor --roster open-weight-heavy

# Run on the current repo (interactive: asks for the change + approves the plan)
node dist/cli.js run

# Or fully unattended
node dist/cli.js run --non-interactive --yes \
  --roster "claude:sonnet,openrouter:qwen/qwen3-coder,oss:gpt-oss:120b" \
  --request "Add a slugify() utility with unit tests"

Artifacts for every run land under .pangloss/runs/<run-id>/ (per-round plan.json, code-outcomes.json, reviews.json, selection.json, summary.md). The winning worktree is kept for inspection.

CLI

Command	Description
pangloss run	Plan → code → review → select (→ revise-loop). The default command.
pangloss agents	List configured rosters and agent presets.
pangloss doctor [--roster …]	Verify roster CLIs are installed/authed; preview invocations.
pangloss models [--filter …]	List OpenRouter model slugs usable as openrouter:<slug>.
pangloss config / pangloss setup	Generate pangloss.config.json / .env.

Key run flags: --roster <name|csv>, --request <text>, --rounds <n>, --keep-worktrees, --timeout <min>, -y/--yes, --non-interactive.

Setup

Requirements: Node 20+, Git 2.5+, and the CLIs for whatever roster you use (claude, codex, cursor-agent, gemini, ollama). Each tool authenticates itself (OAuth/subscription or API key) — Pangloss runs them on the host, so no keys need to be injected anywhere.

OpenRouter (optional, for the openrouter: lanes and the openrouter roster): set OPENROUTER_API_KEY in .env. Discover model slugs with pangloss models.

Local open-weight (the oss: lanes / gptoss): run ollama with the model pulled (e.g. ollama pull gpt-oss:120b). Driven via codex exec --oss.

The agent behavior contract every model obeys inside its worktree lives in .claude/skills/pangloss-worktree/SKILL.md and is injected into every agent's system prompt.

Web apps: a Docker-isolated DB per agent

For targets that need a database (or other services), the compose runtime gives every agent its own stack so they run in parallel — no host-port collisions, and your source tree is never modified. Add a compose block to the manifest pointing at the app's existing docker-compose.yml:

"manifest": {
  "setup": "pnpm install",
  "build": "pnpm build",
  "test":  "pnpm test",           // integration tests hit the DB
  "e2e":   "pnpm test:e2e:ci",    // optional; Playwright self-starts the app
  "compose": {
    "file": "docker-compose.yml",
    "dbService": "db",
    "dbPortBase": 5440,           // agent i gets 5440 + i
    "urlEnv": "DATABASE_URL",
    "urlTemplate": "postgres://test_user:test_password@localhost:{port}/test_db",
    "dbSetup": "pnpm db:rebuild"  // migrate + seed the fresh DB
  }
}

Per agent, Pangloss writes a port-rewritten copy of your compose (the source is untouched), brings it up under a unique project name — docker compose -p pangloss-<run>-<agent> — waits for the DB, runs dbSetup, injects DATABASE_URL, then builds/tests/e2e's against that isolated stack and tears it down (down -v). Agents still run on the host (so their CLI auth just works); only the runtime/DB is containerized.

Run it from a clone you've dedicated to Pangloss (it makes its own worktrees + branches there; your other clones stay free for manual work):

cd ~/projects/msi/technician-app-gamma
node /path/to/pangloss/dist/cli.js run \
  --non-interactive --yes --overnight \
  -c /path/to/pangloss/examples/technician-app.config.json \
  --roster "claude:sonnet,oss:gpt-oss:120b,openrouter:qwen/qwen3-coder" \
  --request "…your feature…"

--overnight is recommended here: heavy e2e + slow local models want no clock. See examples/technician-app.config.json.

Development

yarn build       # tsc
yarn test        # jest
yarn lint        # eslint
yarn typecheck   # tsc --noEmit

License

MIT