Junie CLI

Junie CLI takes a custom endpoint as a JSON profile in $JUNIE_HOME/models/, where baseUrl is the complete endpoint URL and fasterModel decides where its helper traffic lands.

{
  "id": "claude-sonnet-5",
  "baseUrl": "https://api.kunavo.com/v1/chat/completions",
  "displayName": "Kunavo",
  "providerName": "Kunavo",
  "apiType": "OpenAICompletion",
  "apiKey": "${KUNAVO_API_KEY}",
  "fasterModel": {
    "id": "claude-haiku-4-5"
  }
}

Step by step

1. Create a key at /app/keys and copy it — it is shown once. Export it as KUNAVO_API_KEY in the shell you will start Junie from, or put a literal key in the file instead; JetBrains notes that “Values without any $${...} reference are used as-is”.
2. Save the JSON above as $JUNIE_HOME/models/kunavo.json for your account, or .junie/models/kunavo.json to share it with a repository. The filename without .json becomes the profile id — that is what makes the selector below read custom:kunavo.
3. Set id to the model id you actually want. It is described as “The model identifier as expected by the API endpoint”, so a Kunavo slug from the table below goes straight in; the ids are resolved at the endpoint, not inside Junie.
4. Start Junie and pick the profile: junie --model custom:kunavo, or the /model command in an interactive session. Custom models “appear in the model selection list after the built-in providers”.
5. Give it a task that edits a file rather than a greeting. Junie is an agentic client and leans on tool calling for almost everything it does, so a first run that reads and writes something tells you far more than a single reply does.

Checked against JetBrains’ Custom LLMs page for Junie on September 21, 2026. Third-party settings move; if a field name here no longer matches what you see, that page is the authority, not this one.

Verify before you debug the client

One request settles whether a failure is the endpoint, the key, or the configuration file. If this returns JSON, the same base URL and key work in Junie CLI.

# Settles whether a failure is the endpoint, the key, or the client.
curl -sS https://api.kunavo.com/v1/models \
  -H "Authorization: Bearer sk-kn-..."

Which model id to put in the field

Every text model is reachable as a model id — the live list is GET /v1/models, and the catalog with prices is on the models page. Rates are USD per 1M tokens, input / output.

The other two apiType values Kunavo answers

apiType names a request format, not a vendor, and Junie accepts four values: OpenAICompletion, OpenAIResponses, Google and Anthropic. Kunavo serves three routes, so two of the other three values have a matching baseUrl here — and because Junie appends no path, changing the format means changing the URL in the same edit:

1. "apiType": "OpenAICompletion" with https://api.kunavo.com/v1/chat/completions — the profile above, and the one this page recommends because the route is the most widely exercised of the three.
2. "apiType": "Anthropic" with https://api.kunavo.com/v1/messages — Kunavo answers the Messages format on that path.
3. "apiType": "OpenAIResponses" with https://api.kunavo.com/v1/responses.

Which of the three Junie drives best for a given model is exactly the kind of question a runtime test answers and a documentation read does not, so this page does not rank them beyond that first line. Pick one, keep the URL and the apiType in step, and if a key-set profile comes back 401, make the credential explicit rather than guessing at Junie's header: JetBrains documents apiKey only in the negative — “If omitted, requests are sent without an Authorization header” — and never states the scheme used when it is present. extraHeaders is documented as “additional HTTP headers to include in every request” and takes the same $${VAR} references, which makes "extraHeaders": {"Authorization": "Bearer $${KUNAVO_API_KEY}"} the unambiguous form.