TokenVoke

Get gateway endpoints

Return the Tokenera base URLs, the capability->endpoint map (chat/gemini/video/...), and the auth header format. Start here to learn how to call the gateway.

Search documentation

Full-text search the Tokenera user docs (live /api/docs when available, else bundled snapshot). Returns matching docs with snippets.

Read a documentation page

Read a full doc page by slug. Use search_docs or list to find slugs.

List models

List available models with vendor, tags, supported endpoints and USD pricing. Filter by tag/vendor/endpoint/search. Uses the public catalog; set mine=true to list only models your API key can access.

Get model pricing

Live pricing for one or more models (USD per 1M tokens for input/output, cached input, per-call). Filter the same way as list_models.

Estimate request cost

Estimate quota + USD cost for a model given token counts (or call count for per-call models). Pure calculation from live pricing.

Recommend a model

Recommend cost-effective models for a task type (chat, cheap-chat, vision, code, image, video, embedding, translation, roleplay), sorted by price.

Resolve route for a model

Determine the base URL to use and (if a model is given) verify it is sellable on this cluster, returning its endpoint type. This MCP serves one cluster; credentials only work on the cluster they belong to.

Get a ready-to-run call snippet

One-stop model switching: given a model, returns the correct endpoint and ready-to-run curl + python + node snippets (auto base URL + auth + sample body). To switch models, call again with a different model.

Diagnose model access

Check why a model call may fail and return actionable fixes (key prefix, group, availability, Seedance unlimited fallback, empty model list). Runs a quick live check when an API key is provided.

Invoke a chat model (small, validation)

Convenience: make ONE small non-streaming chat call to validate a model works. For real/large traffic, use get_call_snippet and call the gateway directly. Refuses video (Seedance) models. Needs an API key.

Check account balance

Show your remaining balance and usage in USD. Needs a logged-in access token (interim: Authorization access_token + TokenVoke-User header; later: OAuth).

List API keys

List your API keys (masked). Needs a logged-in access token.

Create an API key

Create a new API key (sk-). Optionally cap quota (USD), expiry (days), allowed models and group. Needs a logged-in access token.

Delete an API key

Delete an API key by id. Needs a logged-in access token.

Test a model

Send a tiny chat request to confirm a model is reachable with your key. Refuses video models. Needs an API key.