Skip to main content
Kilo Code is an open source AI coding agent for VS Code and the terminal. Requesty is a built in provider in Kilo Code, so you only need an API key to get started. Using the Requesty integration, you can:
  • Access 600+ models from OpenAI, Anthropic, Google, Mistral, and many other providers through one API key.
  • Track and manage your spend in a single location.
  • Apply fallback policies, load balancing, and latency routing to keep your agent responsive.
  • Keep traffic in the EU with EU routing.

​
Prerequisites

​
VS Code extension

1

Open the Providers settings

Open Kilo Code from the Activity Bar, click the gear icon to open Settings, and go to the Providers tab.
2

Find Requesty

Requesty is not in the Popular providers list. Scroll down, click Show more providers, search for requesty, and select Requesty.Kilo Code Connect provider dialog with Requesty in the search results
3

Paste your Requesty API key

Paste your key into Requesty API key and click Submit. Requesty now shows under Connected providers with an API KEY badge.Kilo Code Connect Requesty dialog with the API key field filled in
4

Pick a model

Click the model selector at the bottom of the chat box and search for a model, for example sonnet 4.6. Pick the entry labelled Requesty, not Kilo Gateway.Kilo Code model picker showing Claude Sonnet 4.6 from RequestyThe Requesty models are Managed Policy names such as claude-sonnet-4-6, gpt-5.5 or deepseek-v4.1-flash. Models ending in @eu (shown as (EU)) run in the EU.
5

Start a task

Type a prompt and press Enter. Kilo Code shows the cost and token usage of the session at the top of the chat.Kilo Code completing a task with Claude Sonnet 4.6 through Requesty
Kilo Code also sends short background requests, such as session titles, to a Small Model. Choose which model handles them under Settings > Models > Small Model.
The extension and the CLI share the same credential store (~/.local/share/kilo/auth.json) and config files, so a key you add in VS Code also works in the terminal.

​
Kilo CLI

Log in once and the key is saved to the same auth.json store:
You can also skip the login and set the key in your shell. Kilo reads REQUESTY_API_KEY automatically:
List the Requesty models and run a prompt:
To set a default model, add it to kilo.json in your project root or ~/.config/kilo/kilo.jsonc for a global config:

​
Use a Routing Policy

A Routing Policy is a named alias that lives on the Requesty side. You change which model (or chain of models) it resolves to in the Requesty UI, and every Kilo Code session picks up the change without config edits. This is the setup we recommend for teams. Kilo Code’s model list only contains Requesty’s built in catalog, so your policies are not listed. Using one without declaring it fails with Model not found: requesty/policy/<name>. Add each policy under provider.requesty.models in kilo.json:
The model key is the policy ID without the requesty/ prefix, as it appears in GET https://router.requesty.ai/v1/models. In VS Code, click Reload at the top of the Kilo Code settings and the policy appears in the model picker under Requesty. The same models block works for any Requesty model that is not in Kilo Code’s list, for example "vertex/gemini-2.5-pro": { "name": "Gemini 2.5 Pro" }.

​
EU routing

To send all traffic through the EU region, override the base URL:
Pair it with the @eu models, for example requesty/claude-sonnet-4-6@eu. See EU Routing for details.

​
Track cost per team, repo, or developer

Add Analytics Headers under options.headers and every Kilo Code request carries them to Requesty:
You can then group spend by Team or Repo in Analytics. The headers are only read by Requesty and are not forwarded to the model provider.

​
Verifying the integration

Run a prompt in Kilo Code, then open Logs or the analytics dashboard. The request shows up within seconds with its model, cost, and token counts.

​
Troubleshooting

Requesty is not one of the Popular providers. Click Show more providers at the bottom of the Providers tab and search for requesty.
Kilo Code does not know your routing policies until you declare them. Add the policy under provider.requesty.models, see Use a Routing Policy.
The key, user, or group hit its spend limit in Requesty. An admin can raise it, see Spend Limits & Rate Limits.
Click Change API key next to Requesty in the Providers tab and paste the key again, or run kilo auth login --provider requesty. If REQUESTY_API_KEY is set in your shell, check that it holds the right key.

​
References

Last modified on October 29, 2026