Skip to content

New: DeepSeek V4.1 Flash Derisked. In chat and on the API.

See modelsWhat's newReport a bug
Open chat
All IDE and SDK guides

JetBrains IDEs

Use the Vani API in JetBrains AI Assistant

Add Vani as an OpenAI-compatible provider in AI Assistant, or as a Junie CLI profile.

URL
https://api.vani.ai/v1
Provider type
OpenAI-compatible
Auth
API key (sent as Bearer)
Example model
openai/gpt-6-sol

Checked against the vendor's documentation on 2 October 2026. Need a key? Create one in the API console.

On this pageBefore you start

Before you start

  • The AI Assistant plugin (IDE 2023.3 or later; Community editions 2024.1.1 or later). Your own key works without a JetBrains AI subscription; some features need one.
  • Code completion and next edit suggestions are not powered by third-party providers: leave AI Completion on JetBrains.
  • Junie inside the IDE does not use custom endpoints; Junie CLI does (below).
  • If your organization manages AI through JetBrains IDE Services, an administrator can restrict providers.

AI Assistant

  1. Open Settings | Tools | AI Assistant | Providers & API keys.
  2. In Third-party AI providers choose OpenAI-compatible. URL: https://api.vani.ai/v1. API Key: your Vani key.
  3. Turn on Tool calling (lets the model call MCP tools; Vani supports function calling), click Test Connection, then Apply.
  4. In Models Assignment set Core features to openai/gpt-6-sol and Instant helpers to openai/gpt-6-sol-fast (or deepseek/deepseek-v4.1-flash), and set the context window to the model's real window. Apply.
  5. In AI Chat, open the model menu and choose the Vani model.
Text

Providers & API keys

Provider        : OpenAI-compatible
URL             : https://api.vani.ai/v1
API Key         : vani_YOUR_API_KEY
Tool calling    : on
Core features   : openai/gpt-6-sol
Instant helpers : openai/gpt-6-sol-fast

Note

JetBrains documents the field as the provider's API endpoint; the /v1 base form shown here is the OpenAI-compatible convention used by provider guides. Test Connection confirms it.

Junie CLI

  1. Create a profile at $JUNIE_HOME/models/vani.json (or .junie/models/vani.json in a project).
  2. baseUrl is the complete endpoint URL: Junie does not append a path.
  3. Export VANI_API_KEY, then run junie --model custom:vani.
JSON

$JUNIE_HOME/models/vani.json

{
  "baseUrl": "https://api.vani.ai/v1/chat/completions",
  "id": "openai/gpt-6-sol",
  "apiType": "OpenAICompletion",
  "apiKey": "${VANI_API_KEY}",
  "displayName": "Vani GPT-6 Sol",
  "providerName": "Vani",
  "fasterModel": {
    "id": "openai/gpt-6-sol-fast"
  }
}
Shell

Run

export VANI_API_KEY="vani_YOUR_API_KEY"
junie --model custom:vani

Troubleshooting

Test Connection fails or returns 404
Use https://api.vani.ai/v1 and the OpenAI-compatible provider (not OpenAI). Check the URL with curl https://api.vani.ai/v1/models -H "Authorization: Bearer $VANI_API_KEY".
Some features still use JetBrains models
Assign the Vani models in Models Assignment. Code completion and next edit suggestions never use third-party providers.
MCP tools are never called
Turn on Tool calling for the provider and click Apply.
Junie CLI: the profile fails to load, or 404
Export VANI_API_KEY before starting Junie (a missing variable stops the profile from loading) and use the full https://api.vani.ai/v1/chat/completions as baseUrl.
404 model_not_available, or "model not found"
Use the exact vendor-prefixed id, for example openai/gpt-6-sol. List the ids your key can use with curl https://api.vani.ai/v1/models -H "Authorization: Bearer $VANI_API_KEY".

Still stuck? Check service status or the API error reference.

Sources