Generate text using an LLM. Routes to the best available provider with automatic fallback. Supports three-part prompts (system/context/instruction) for improved quality.

Store an API key in the encrypted credential vault. Upserts by (provider, keyName, project).

List all stored credentials with masked values. Optionally filter by project.

Delete a stored credential by its ID.

List all available models across registered providers.

Store an auth file (e.g. auth.json) in the encrypted vault. Upserts by (provider, fileName, project).

List all stored auth files (metadata only). Optionally filter by project.

Delete a stored auth file by its ID.

List all provider groups for load balancing.

Create a new provider group for load balancing.

Delete a provider group by its ID.

Configure circuit breaker settings. Updates thresholds and backoff for all breakers.

Get circuit breaker stats for all providers. Shows state, failures, successes, cooldown.

Get cost/usage summary. Returns total requests, tokens, cost, with optional breakdown by provider, model, project, hour, or day.

Query individual usage records with filters. Returns raw usage log entries.

Search code semantically. Finds functions, classes, and blocks matching a query using keyword + fuzzy matching. Optionally follows imports for related code.

Index a codebase directory for semantic code search. Scans files, extracts functions/classes/blocks, and builds an in-memory search index.

CRDT-based shared state for multi-agent collaboration. Supports conflict-free read/write/merge with G-Counter (token tracking), LWW-Register (agent status), and OR-Set (shared findings).

List pending approval requests.

Approve a pending request by ID.

Deny a pending request by ID.

Generate text using a local LLM (Ollama/LM Studio) for offloadable tasks. Falls back to cloud provider if local LLM is unavailable.

Discover local LLM models and enrich them with HuggingFace metadata. Returns enriched model list with capabilities and recommended tasks.

Divide a long conversation into navigable pages. Use this when conversation exceeds safe context limits.

Get a specific page from a paginated conversation

Get a page with surrounding context pages. Use this for reading with context.

Navigate to next, previous, first, or last page

Get info about a paginated conversation: total pages, total tokens, etc.

Find pages relevant to a query using keyword matching

Check if conversation needs compaction for given model context limit