Skip to contentKumoDocs
Documentation

One key for every model

Change the base URL and the key — the rest of your code stays as it is. The platform itself publishes which models are available, what they can do, and what they cost.

curl https://api.kumorouter.com/v1/chat/completions \
  -H "Authorization: Bearer $KUMO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "<model>",
    "max_tokens": 128,
    "messages": [{ "role": "user", "content": "Explain tokens in one line." }]
  }'

Integrations

IntegrationsPoint the tool you already use at the gateway: a base URL and a key. A recipe for every client, step-by-step guides, and a prompt that lets your agent do the setup.Connect your toolPick your client and get numbered steps with ready settings files — install, connect, launch, choose a model, turn on thinking — with the gateway address already in place.Claude CodeGateway environment variables and Claude Code runs through Kumo. Plus the settings.json route, classifier billing, key precedence, and a one-command check.Codex CLIA provider in config.toml, the key from an environment variable — Codex CLI runs through the gateway without changing a single habit of yours.CursorYour own key and an overridden base URL in Cursor's model settings, and the editor's chat runs through the gateway. Plus an honest list of what does not.VS Code extensionsCline, Roo Code and Continue all take the same OpenAI-compatible provider: base URL, key, model ID. Here is where those fields live in each of them.Other toolsAider, Zed, OpenClaw — and the way to connect a tool that is not on this list: hand the connection details to a coding agent and let it do the setup.SDKsThe official OpenAI and Anthropic SDKs, Python and Node: install, a client with the gateway's base URL and the key from the environment, one call. Your code stays your code.Kumo Usage CLI BarA spend-and-remainder panel under the Claude Code and Codex prompt — one command to install, it finds your key itself, and it shows what is left: a package, the balance, or your own ceiling.

The gateway

Chat CompletionsThe POST /v1/chat/completions call — request members, conversation roles, the response shape, and pointers to streaming, tools and structured output.ResponsesThe POST /v1/responses call — how it differs from Chat Completions, the request members, the typed output items and the stream's terminal event.MessagesThe Anthropic Messages dialect — the bare base URL, two carriers for the key, request members, reply blocks, stream events, and a token count that spends no quota.EmbeddingsThe POST /v1/embeddings call — a batch of inputs, one base64 vector per input, and what the call costs.ImagesThe POST /v1/images/generations call — the prompt, how many images, the size, the response shape, and why generation is counted in images rather than tokens.DecisionsThe POST /v1/systemone call — a state and questions in, answers as probabilities out, and what the call costs.StreamingSSE across the gateway's three dialects — how to turn a stream on, which events arrive, what marks a whole answer, and why a call is never moved after the first byte.ToolsFunction calling across the gateway's three dialects — declaring a tool, the call-and-result round, and why tool support is proven per model.Structured outputAn answer to a named JSON Schema — what format the API takes, what strict means, how it is spelled on the three dialects, and why support is proven separately for every model.ErrorsThe envelope a refusal arrives in, the status codes the gateway answers with, and the two behaviors every client must handle.Rate limitsThe ceilings a call is measured against, what a call that reaches one gets back, and how a higher ceiling is asked for.API referenceEvery operation of the Kumo gateway, generated from the published API description: the address, how it is authenticated, what the request carries and what comes back.

Account and billing

Running an agent? The whole of this documentation is available as Markdown at/llms-full.txt.