Arcee AI Provider Setup and Model Selection
Learn how to install and configure the Arcee AI plugin for OpenClaw, including API key setup and model selection. This guide is for developers using Trinity models directly or via OpenRouter.
Read this when
- You want to use Arcee AI with OpenClaw
- You need the API key env var or CLI auth choice
Arcee AI offers the Trinity line of mixture-of-experts models via an API that follows OpenAI conventions. Every Trinity model carries an Apache 2.0 license. Arcee comes as an official OpenClaw plugin rather than part of the core distribution, so you must install it before you can start using it.
You can reach Arcee models either directly through the Arcee platform or by going through OpenRouter.
| Property | Value |
|---|---|
| Provider | arcee |
| Auth | ARCEEAI_API_KEY (direct) or OPENROUTER_API_KEY (via OpenRouter) |
| API | OpenAI-compatible |
| Base URL | https://api.arcee.ai/api/v1 (direct) or https://openrouter.ai/api/v1 (OpenRouter) |
Install plugin
openclaw plugins install @openclaw/arcee-provider
openclaw gateway restart
Getting started
Direct (Arcee platform)
Get an API key
Head over to Arcee AI to generate an API key.
Run onboarding
openclaw onboard --auth-choice arceeai-api-key
Set a default model
{
agents: {
defaults: {
model: { primary: "arcee/trinity-large-thinking" },
},
},
}
Via OpenRouter
Get an API key
Generate an API key on OpenRouter.
Run onboarding
openclaw onboard --auth-choice arceeai-openrouter
Set a default model
{
agents: {
defaults: {
model: { primary: "arcee/trinity-large-thinking" },
},
},
}
Both the direct and OpenRouter paths accept the same set of model refs.
Non-interactive setup
Direct (Arcee platform)
openclaw onboard --non-interactive --accept-risk --skip-health \
--mode local \
--auth-choice arceeai-api-key \
--arceeai-api-key "$ARCEEAI_API_KEY"
Via OpenRouter
openclaw onboard --non-interactive --accept-risk --skip-health \
--mode local \
--auth-choice arceeai-openrouter \
--openrouter-api-key "$OPENROUTER_API_KEY"
Direct Arcee catalog
| Model ref | Name | Input | Context | Max output | Cost (in/out per 1M) | Tools | Notes |
|---|---|---|---|---|---|---|---|
arcee/trinity-large-thinking | Trinity Large Thinking | text | 256K | 80K | $0.25 / $0.90 | No | Default model; extended thinking |
arcee/trinity-large-preview | Trinity Large Preview | text | 128K | 16K | $0.25 / $1.00 | Yes | General-purpose; 400B params, 13B active |
arcee/trinity-mini | Trinity Mini 26B | text | 128K | 80K | $0.045 / $0.15 | Yes | Fast and cost-efficient; function calling |
Tip
The onboarding preset designates
arcee/trinity-large-thinkingas the model used by default.
OpenRouter catalog
During OpenRouter onboarding, arcee/trinity-large-preview and arcee/trinity-large-thinking are exposed. OpenClaw stores those provider-qualified model refs in its configuration and passes OpenRouter's canonical arcee-ai/* runtime ids along. OpenRouter no longer serves Trinity Mini; to use that model, go through the Arcee API directly.
Supported features
| Feature | Supported |
|---|---|
| Streaming | Yes |
| Tool use / function calling | Yes (Trinity Mini, Trinity Large Preview) |
| Structured output (JSON mode and JSON schema) | Yes |
| Extended thinking | Yes (Trinity Large Thinking; tools disabled) |
Environment note
When the Gateway runs as a daemon (launchd/systemd), confirm that ARCEEAI_API_KEY
(or OPENROUTER_API_KEY) is accessible to that process, for instance in
~/.openclaw/.env or through env.shellEnv.
OpenRouter routing
OpenRouter relies on the same arcee/trinity-large-thinking OpenClaw model ref.
OpenClaw handles routing using the canonical arcee-ai/trinity-large-thinking
OpenRouter runtime id. For OpenRouter-specific configuration
details, check the
OpenRouter provider docs.
Related
-
OpenRouter, Reach Arcee models plus a wide range of others with one API key.
-
Model selection, Picking providers, model refs, and failover behavior.