Coding tools
JetBrains IDEs
AI Assistant ships an OpenAI-Compatible provider first-party, across every JetBrains IDE — no plugin required. Chat and inline completion are configured separately.
Wire format
OpenAI format
Base URL
https://api.xkiro.com/v1
Configured by
Settings | Tools | AI Assistant
Add xKiro as a provider#
Open Settings | Tools | AI Assistant | Providers & API keys, find the Third-party AI providers section, and select OpenAI-Compatible.
| Field | Type | Description |
|---|---|---|
API Key | string | Specify the API key to access the models provided by the selected third-party AI provider. |
URL | string | Specify the URL of the provider's API endpoint — https://api.xkiro.com/v1. A Test Connection button next to the field checks it immediately. |
Tool calling | toggle | Specify whether the model supports calling tools configured through the Model Context Protocol (MCP). |
Assign models#
Once the provider is added, assign its models under Models Assignment:
| Field | Type | Description |
|---|---|---|
Core features | model | The model AI Assistant uses for its main chat and agent features. |
Instant helpers | model | The model used for smaller, lower-latency helper actions. |
Context window | numberdefault: 64,000 | Tokens of context the assigned model is allowed to use. Match this to the real model — see GET /v1/models. |
Chat and inline completion are two separate configurations
Adding xKiro as a provider and assigning it above only covers chat and agent features. Inline completion (the ghost text as you type) reads its own, separate provider setting — configuring one does not enable the other.
Inline completion#
In the AI Completion section, select OpenAI Compatible and fill in:
| Field | Type | Description |
|---|---|---|
API key | string | Your xKiro API key. |
Base URL | string | https://api.xkiro.com/v1. Also has its own Test Connection button. |
Model | string | The model ID to use for completions. |
Model context | number | Context window for the completion model. |
Max output tokens | number | Output limit for a single completion. |
Prompt schema | select | The prompt format the completion model expects. |
MCP tool calling on a custom provider — unconfirmed
JetBrains' own docs say: "Currently, AI Assistant does not support invoking tools from configured MCP servers when using local models." That statement is about local models specifically; we could not confirm whether it also applies to an OpenAI-Compatible provider like xKiro, so treat MCP tool calling here as unconfirmed rather than assuming either way.
No JetBrains AI subscription required#
Basic use of a custom OpenAI-Compatible provider does not require a JetBrains AI subscription. A subscription can still be activated alongside it as a fallback, for features the custom provider cannot serve.
Plugins are a fallback, not the primary path#
Because AI Assistant supports custom OpenAI-compatible providers first-party, plugins like Continue, ProxyAI or CodeGPT are a fallback for JetBrains IDEs, not the recommended way to reach xKiro.
If it does not work#
Send one request by hand before touching JetBrains IDEs's settings again. A reply proves the URL and key are both fine, which means whatever is left is in the tool's own configuration — and that is a much smaller place to look.
curl https://api.xkiro.com/v1/chat/completions \
-H "Authorization: Bearer $XKIRO_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"openai/gpt-5.6-sol","messages":[{"role":"user","content":"ping"}],"max_tokens":8}'An error naming an unknown model still counts as a success here: the request was authenticated before the model was rejected. Common failures and what each one means are on the overview page.