Integration Guide

Forge + OpenClaw.
Private AI agents at flat rate.

Run your OpenClaw agents on self-hosted infrastructure with predictable costs. One config change — no code modifications, no per-token billing, full data sovereignty.

Why it matters

OpenClaw agents run 24/7. Per-token pricing doesn't.

Flat rate for always-on agents

OpenClaw's persistent daemon and heartbeat scheduler make continuous API calls. Forge's flat monthly pricing makes always-on agents economically viable instead of a billing surprise.

Sovereign inference for private agents

OpenClaw accesses your email, calendars, files, and messaging platforms. Sovereign mode ensures the LLM processing those sensitive requests stays on infrastructure you control.

Vision + text from one endpoint

OpenClaw supports vision tasks and browser automation. Forge's self-hosted models handle text, images, and documents through a single API — no provider juggling needed.

Setup in 2 minutes

One config block. That's it.

OpenClaw natively supports any OpenAI-compatible API endpoint. Add Forge as a custom provider in your agent's configuration — no plugins, no code changes.

Get your Forge API key

Add the provider config

Add the Forge provider block to your OpenClaw config file (see code example).

Start your agent

Your OpenClaw agent now routes all inference through Forge. Smart routing picks the best model automatically.

openclaw provider config

// ~/.openclaw/openclaw.json
{
  "models": {
    "providers": {
      "forge": {
        "baseUrl": "https://forge-api.lanaai.io/v1",
        "apiKey": "${FORGE_API_KEY}",
        "api": "openai-completions",
        "models": [
          { "id": "auto", "name": "Forge Auto" },
          { "id": "fast", "name": "Forge Fast" },
          { "id": "reasoning", "name": "Forge Reasoning" }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": { "primary": "forge/auto" }
    }
  }
}

Sovereign mode (add headers)

// Add headers to keep all data on private infrastructure
"forge": {
  "baseUrl": "https://forge-api.lanaai.io/v1",
  "apiKey": "${FORGE_API_KEY}",
  "api": "openai-completions",
  "headers": {
    "X-Sovereign": "true"
  },
  "models": [
    { "id": "auto", "name": "Forge Auto" }
  ]
}

SOUL.md snippet

# In your agent's SOUL.md
I use the Forge provider for all
inference. When handling sensitive data
(emails, contracts, medical records),
I always use the forge-sovereign provider
to keep data on private infrastructure.

Quick install

Paste this into your terminal.

This sets your Forge API key as an environment variable and configures OpenClaw to use Forge as its default provider.

Step 1 — Set your API key

# Add to ~/.bashrc, ~/.zshrc, or ~/.profile
export FORGE_API_KEY="your-key-here"

Step 2 — Add Forge provider to OpenClaw

// Add to ~/.openclaw/openclaw.json
{
  "models": {
    "providers": {
      "forge": {
        "baseUrl": "https://forge-api.lanaai.io/v1",
        "apiKey": "${FORGE_API_KEY}",
        "api": "openai-completions",
        "models": [{ "id": "auto", "name": "Forge Auto" }]
      }
    }
  },
  "agents": { "defaults": { "model": { "primary": "forge/auto" } } }
}

Step 3 — Verify it works

# Test your connection
curl -s https://forge-api.lanaai.io/v1/chat/completions \
  -H "Authorization: Bearer $FORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{"role": "user", "content": "Hello from OpenClaw"}],
    "max_tokens": 20
  }'

# You should see a JSON response with the model's reply.
# If it works here, it works in OpenClaw.

Don't have an API key yet? Sign up free — 1M tokens on us.

What Forge unlocks

What your OpenClaw agent can do with Forge.

Without Forge

Per-token billing adds up fast with always-on agents
Your emails, docs, and messages sent to OpenAI/Anthropic servers
Locked to one provider per config — switching means reconfiguring
No audit trail of what your agent accessed or sent

With Forge

Flat rate from $49/mo — run your agent 24/7 without bill anxiety
Sovereign mode keeps all data on private infrastructure
22+ models from one endpoint — auto-routing picks the best one
Full compliance audit trail for every request your agent makes

Use cases

What you can build with Forge + OpenClaw.

Private executive assistant

An OpenClaw agent connected to your email, calendar, and Slack — with sovereign mode ensuring that meeting notes, contracts, and internal communications never leave your infrastructure.

Compliance-safe document processing

Process legal documents, medical records, or financial reports through an OpenClaw agent backed by Forge's sovereign inference. Full audit trail of every document touched.

Multi-platform support bot

OpenClaw connects to 12+ messaging platforms. Power it with Forge to handle customer support across WhatsApp, Slack, Teams, and Discord from a single agent with cost-optimized routing.

Robotics and edge agents

Running OpenClaw on a Jetson, RUBIK Pi, or humanoid robot? Forge provides reliable self-hosted inference without depending on cloud provider uptime or internet connectivity.

Run your agent on your terms.

Get a Forge API key and connect your OpenClaw agent in under 2 minutes.

Get Started API Docs

Forge + OpenClaw. Private AI agents at flat rate.