Kilo Code on 600+ models, from the provider list
Requesty ships as a built-in provider in Kilo Code, in the VS Code extension and the CLI. Paste a key, pick a model, and every request is logged, costed and held to the spend cap on that key.
{ "$schema": "https://app.kilo.ai/config.json", // a Requesty model is requesty/ plus its ID "model": "requesty/claude-sonnet-5-5" }
Set up Kilo Code in three steps
Requesty sits in Kilo's provider list, so there is no base URL to type. The extension and the CLI share one credential store and the same config files, so a key added in one works in the other.
- 01Create a key
Make an API key on the Requesty API Keys page. A key per developer gives each one its own spend cap.
- 02Add Requesty as a provider
In the Kilo Code extension, open Settings, then the Providers tab. Click Show more providers, search for Requesty, paste your key and click Submit.
- 03Pick a model
In the model selector under the chat box, choose the entry labelled Requesty, not Kilo Gateway. Session cost and tokens show at the top of the chat.
# log in once, or export REQUESTY_API_KEY instead kilo auth login --provider requesty # list the models your key can use, then run one kilo models requesty kilo run -m requesty/claude-sonnet-5-5 "Summarize this repo"
Use any Requesty model in Kilo Code
In the CLI and in kilo.json, a model is requesty/ followed by its Requesty ID. An @eu suffix runs the model in the EU, and a routing policy is requesty/policy/<name>.
A policy, or a model missing from Kilo's list, also needs an entry under provider.requesty.models in kilo.json. The guide shows the block.
of provider failures on the gateway in April 2026 were 429s. Throttling, not downtime, is what stops an agent mid-task.
eventual success on a routing policy in April 2026, against 85.01% for callers pinned to one provider.
Keep Kilo working when a provider throttles
Long agent runs make hundreds of calls, and one unretried 429 can end a task. Point Kilo at a routing policy and a throttled request retries on the next model in its chain. Team and repo headers split the bill, and a monthly cap on the key bounds it.
- Routing policiesrequesty/policy/<name> retries on another model when the first fails
- Cost per team or repoX-Requesty-Team and X-Requesty-Repo headers in kilo.json
- Spend capsA monthly limit per key, user or team
Share of April 2026 requests where the upstream provider returned a failure. Throttling, not downtime, is what stops an agent mid-task. See the data note.
Kilo Code with Requesty, answered
How do I use Kilo Code with my own API key?
Requesty is a built-in provider. In the Kilo Code extension, open Settings, then Providers, add Requesty and paste your Requesty API key. One key then covers 600+ models from every major provider, billed at the provider price plus 5%.
How do I configure Requesty in the Kilo CLI?
Install it with npm install -g @kilocode/cli, then run kilo auth login --provider requesty or export REQUESTY_API_KEY. kilo models requesty lists your models. For a default, set model to requesty/claude-sonnet-5-5 in kilo.json at the project root or in ~/.config/kilo/kilo.jsonc.
Does Kilo Code use OpenRouter?
Kilo has its own Kilo Gateway and also lists other gateways as providers, including OpenRouter and Requesty. With Requesty selected, requests go through your Requesty key and show up in your Requesty dashboard.
How do I fix "rate limit exceeded" in Kilo Code?
The error comes from the model provider throttling the request. Requesty does not cap requests per minute, and a fallback policy retries a throttled request on the next model in its chain, so the task keeps going.
Which models can I use in Kilo Code with Requesty?
Any model your key is approved for in the Requesty Model Library: Claude, GPT, Gemini, DeepSeek, Kimi, GLM and more. Prefix the ID with requesty/ when you set it in kilo.json. A model missing from Kilo's list goes under provider.requesty.models, then click Reload in settings.
Can I keep Kilo Code traffic in the EU?
Yes. Set provider.requesty.options.baseURL to https://router.eu.requesty.ai/v1 in kilo.json and pick a model with the @eu suffix, such as requesty/claude-sonnet-5-5@eu.
Are there free models for Kilo Code?
Requesty includes 200 requests a day on free models with no credit card, and the same key works for paid models when you need more.
Can I see what Kilo Code costs per model?
Yes. Every Kilo request appears in the Requesty analytics dashboard with its model, tokens, cost and latency, and you can set a monthly spend cap on the key Kilo uses.
Guides for the other agents you run
The same key and the same dashboard work across every harness, so spend and logs stay in one place.
Run Kilo Code on Requesty
600+ models, a spend cap on every key, and every session on the record.