anna docs

Tiered Models

anna supports two model tiers for different workloads:

Tier	Use Case
`strong`	Heavy reasoning, complex tasks
`fast`	Quick responses, simple queries

Each tier falls back independently to model (top-level default) when not set.

model: claude-sonnet-4-6
model_strong: claude-opus-4-6
model_fast: claude-haiku-4-5

CLI Commands

anna models             # List available models (alias for list)
anna models list        # List all models grouped by provider
anna models update      # Fetch models from provider APIs and update cache
anna models current     # Show active provider/model
anna models set <p/m>   # Switch model (e.g. anna models set openai/gpt-4o)
anna models search <q>  # Search models by name

Model Cache

anna models update queries all configured provider APIs and saves results to ~/.anna/cache/models.json. The cache is used by list, search, and the Telegram model picker.

If no cache exists, only models explicitly listed in the config are shown.

Provider Setup

Anthropic

providers:
  anthropic:
    api_key: "sk-..."

Or: export ANTHROPIC_API_KEY="sk-..."

OpenAI

providers:
  openai:
    api_key: "sk-..."
    base_url: "https://api.openai.com/v1"  # optional

Or: export OPENAI_API_KEY="sk-..." and optionally export OPENAI_BASE_URL="..."

OpenAI-Compatible (Responses API)

For services like Perplexity, Together.ai, or any OpenAI-compatible API:

providers:
  openai-response:
    api_key: "sk-..."
    base_url: "https://api.perplexity.ai"

Uses the same OPENAI_API_KEY and OPENAI_BASE_URL env vars as the openai provider.

Runtime Switching

Models can be switched at runtime:

CLI: Via in-chat /model command
Telegram: Via inline keyboard model picker
Config: anna models set provider/model persists the selection to config

Model Metadata

Providers can include detailed model metadata in config:

providers:
  anthropic:
    api_key: "sk-..."
    models:
      - id: claude-sonnet-4-6
        reasoning: false
        input: ["text", "image"]
        context_window: 200000
        max_tokens: 8192
        headers: {}
        cost:
          input: 3.0
          output: 15.0
          cache_read: 0.3
          cache_write: 3.75

This metadata is used for model resolution and display. When not provided, models are constructed with minimal defaults.

Model Management