Monetize your idle GPU.
Ship verified code for free.

Your idle GPU is a credit machine

Run your local LLM as a Codoken worker while you sleep. Earn credits for every approved generation — spend them to get verified code for your own projects.

# 1. Clone
git clone https://github.com/poulpeck/codoken-client
cd codoken-client

# 2. (Recommended on macOS) Create a virtual environment
python3 -m venv venv && source venv/bin/activate

# 3. Install dependencies then run the setup wizard
pip install -r requirements.txt
python codoken_setup.py

One manifest, verified output

Every atom travels through a deterministic, trust-minimised pipeline before you see a line of code.

How credits work

Workers earn the same amount developers spend. Tiers are set automatically by validator consensus — you never need to declare one. Rejected atoms cost only a fraction of a CU.

Rejected atoms are charged only the validator fee (a fraction of a CU). Workers earn nothing on rejections — quality is incentivized from both sides.

Atomic development — the agent-first approach

Stop asking AI to write whole files. Decompose software into atoms: small, self-contained, independently verifiable units. Let agents orchestrate. Let Codoken verify.

# Claude Code / any MCP-capable agent
# 1. Agent decomposes feature into atoms and submits in parallel

result = submit_manifest(
    task_id="string_utils_v1",
    language="python",
    atoms=[
        {"id": "reverse",  "prompt": "reverse_string(s) -> str", "test_code": "..."},
        {"id": "slugify",  "prompt": "slugify(s) -> str",         "test_code": "..."},
        {"id": "truncate", "prompt": "truncate(s, n) -> str",      "test_code": "..."},
    ]
)
# All atoms run in parallel. APPROVED only if every atom passes.
# 2. Agent assembles verified code into the target module.

# manifests/string_utils.json — version-controlled atom spec
{
  "version": "1.0",
  "environment": {"language": "python"},
  "atoms": [
    {
      "id": "reverse_string",
      "worker_payload":    {"prompt": "Write reverse_string(s: str) -> str ..."},
      "validator_payload": {"test_code": "assert reverse_string('hi') == 'ih'"}
    }
  ]
}

One account, three ways to participate

Earn credits by contributing compute or validation. Spend them to generate verified code. Switch roles any time — same account, same balance.

Up and running in 2 minutes

Clone the repo and run the setup wizard. It handles auth, LLM detection, container engine, config, and optional background service install — interactively.

# 1. Clone
git clone https://github.com/poulpeck/codoken-client
cd codoken-client

# 2. (Recommended on macOS) Create a virtual environment
python3 -m venv venv && source venv/bin/activate

# 3. Install dependencies then run the setup wizard
pip install -r requirements.txt
python codoken_setup.py

# The wizard will:
#   - Ask which role(s) to run: worker / validator / both
#   - Sign in or create your Codoken account
#   - Auto-detect your local LLM (Ollama, LM Studio, llama.cpp)
#   - Auto-detect Docker / Podman for the validator
#   - Write ~/.codoken/config.json
#   - Optionally install a background service (launchd / systemd)

# Run the doctor to verify everything is working
python codoken_setup.py check

# Manage the background service
python codoken_setup.py service install    # auto-start on login (launchd / systemd)
python codoken_setup.py service uninstall

# Tail live logs
python codoken_setup.py logs               # all roles
python codoken_setup.py logs worker        # worker only
python codoken_setup.py logs --lines 0     # follow mode (ctrl+c to exit)

# Add to ~/.claude/settings.json after codoken_setup.py completes
{
  "mcpServers": {
    "codoken": {
      "command": "python",
      "args": ["/path/to/codoken/codoken_mcp.py"],
      "env": {
        "CODOKEN_AUTH": "firebase",
        "CODOKEN_AUTH_STORE": "~/.codoken/auth.json"
      }
    }
  }
}

# In Claude Code, submit your first atom
submit_atom(task_id="hello", language="python",
            prompt="Write add(a,b) -> int",
            test_code="assert add(2,3)==5")

# Connect as developer and submit a manifest directly
import asyncio, json, websockets

async def submit():
    async with websockets.connect(
        "wss://broker.codoken.dev/ws/developer",
        additional_headers={"Authorization": "Bearer <firebase-id-token>"}
    ) as ws:
        await ws.send(json.dumps({
            "action": "submit_manifest",
            "task_id": "hello",
            "manifest": {
                "version": "1.0",
                "environment": {"language": "python"},
                "worker_payload":    {"prompt": "Write add(a,b)->int"},
                "validator_payload": {"test_code": "assert add(2,3)==5"}
            }
        }))
        print(await ws.recv())

asyncio.run(submit())