6
Harnesses
Harnesses and models
An agent harness drives one model. yaaf is the harness around the harnesses: put each person on the chart on the harness and model that suits the job, and the whole team still answers to one budget and one set of rules.
6
Harnesses
23
Providers
26
Models in the catalog
1
Budget per project
§ Two levels
Claude Code, Codex and the rest decide how a model is driven. They never decide what it may do. That lives in yaaf's server, the same for every harness.
A harness that can't report what it spent while it's spending it isn't let in. Every one runs inside yaaf's sandbox, on yaaf's meter.
§ Model classes
A role asks for frontier, balanced or fast. Each provider says which of its models that means, so the same team runs on whichever provider you pay.
The most capable model the provider sells. For the CEO and the hardest seats.
The workhorse. Most of a team runs here.
Cheap and quick, for the seats that read and sort.
Pin a seat to an exact model whenever you want. If a provider keeps failing, a project can name a fallback credential: its class seats move there until the provider recovers, and back. Pinned seats never move.
§ Why
Put your researcher on Claude and its critic on GPT. Disagreement gets real.
The rules never depended on the model. Now the metering doesn't either.
§ Harnesses
| Harness | Type | Pays with | Version |
|---|---|---|---|
| Claude Agent SDK | SDK | Anthropic API key · Claude subscription | latest |
| Claude Code CLI | CLI | Anthropic API key · Claude subscription | latest |
| yaaf agent loop | API loop | Anthropic API key · OpenAI API key · Gemini API key · Vertex AI service account · Azure OpenAI API key · AWS Bedrock API key · DeepSeek API key · Mistral API key · xAI (Grok) API key · OpenRouter API key · Groq API key · Together AI API key · Fireworks AI API key · Cerebras API key · Cohere API key · NVIDIA NIM API key · Hugging Face API key · Z.ai (GLM) API key · Moonshot (Kimi) API key · Qwen (Alibaba Model Studio) API key · Ollama (local) · LM Studio (local) · Custom endpoint key | — |
| Codex CLI | CLI | OpenAI API key · ChatGPT plan (Codex login) | 0.162.1 |
| Gemini CLI | CLI | Gemini API key · Google account login (Gemini CLI) | 0.63.0 |
| OpenCode CLI | CLI | OpenAI API key · DeepSeek API key · Mistral API key · xAI (Grok) API key · Groq API key · Together AI API key · Fireworks AI API key · Cerebras API key · Cohere API key · NVIDIA NIM API key · Hugging Face API key · Z.ai (GLM) API key · Moonshot (Kimi) API key · Qwen (Alibaba Model Studio) API key · Ollama (local) · LM Studio (local) · Custom endpoint key | 1.19.0 |
Every person on the org chart runs in one of these. Mix them freely on one team.
§ Providers
| Provider | Credentials | Through |
|---|---|---|
| Anthropic | Anthropic API key · Claude subscription | Claude Agent SDK · Claude Code CLI · yaaf agent loop |
| OpenAI | OpenAI API key · ChatGPT plan (Codex login) | yaaf agent loop · Codex CLI · OpenCode CLI |
| Google (Gemini) | Gemini API key · Google account login (Gemini CLI) | yaaf agent loop · Gemini CLI |
| Google Vertex AI | Vertex AI service account | yaaf agent loop |
| Azure OpenAI | Azure OpenAI API key | yaaf agent loop |
| AWS Bedrock | AWS Bedrock API key | yaaf agent loop |
| DeepSeek | DeepSeek API key | yaaf agent loop · OpenCode CLI |
| Mistral | Mistral API key | yaaf agent loop · OpenCode CLI |
| xAI (Grok) | xAI (Grok) API key | yaaf agent loop · OpenCode CLI |
| OpenRouter | OpenRouter API key | yaaf agent loop |
| Groq | Groq API key | yaaf agent loop · OpenCode CLI |
| Together AI | Together AI API key | yaaf agent loop · OpenCode CLI |
| Fireworks AI | Fireworks AI API key | yaaf agent loop · OpenCode CLI |
| Cerebras | Cerebras API key | yaaf agent loop · OpenCode CLI |
| Cohere | Cohere API key | yaaf agent loop · OpenCode CLI |
| NVIDIA NIM | NVIDIA NIM API key | yaaf agent loop · OpenCode CLI |
| Hugging Face | Hugging Face API key | yaaf agent loop · OpenCode CLI |
| Z.ai (GLM) | Z.ai (GLM) API key | yaaf agent loop · OpenCode CLI |
| Moonshot (Kimi) | Moonshot (Kimi) API key | yaaf agent loop · OpenCode CLI |
| Qwen (Alibaba Model Studio) | Qwen (Alibaba Model Studio) API key | yaaf agent loop · OpenCode CLI |
| Ollama | Ollama (local) | yaaf agent loop · OpenCode CLI |
| LM Studio | LM Studio (local) | yaaf agent loop · OpenCode CLI |
| Custom endpoint (OpenAI-compatible) | Custom endpoint key | yaaf agent loop · OpenCode CLI |
6 models ship with prices, all from Anthropic. Price any other model in Settings. Zero is fine for your own server.
§ 8 rules
The seat's harness must accept the credential.
The harness must honour the role's tool permissions.
If yaaf prices the run, the model needs a price.
One project, one unit: dollars or tokens, never both.
The project needs a credential for the seat's provider.
The model must support the effort level asked for.
The CEO's harness must be able to hold a conversation.
A seat that names a class needs a model for that class on its provider.
§ Status
yaaf smoke checks one on yours.