Can Kilo Code use camelStream as a provider?

Yes. Kilo Code documents custom OpenAI-compatible providers, and camelStream speaks that API. Add the provider in settings, and the model picker discovers Stream's model on its own.

Why does the model picker show only one model?

That's expected, not an error. Stream publishes a single model ID, auto, and routes each request to a frontier model from the fleet behind it. Every model in the fleet scores 70% or higher on Terminal-Bench 2.1, or 50 or higher on the AA Intelligence Index, and you always get its newest public release.

Do I lose Kilo Code's built-in models or my credits?

No. A custom provider sits alongside the built-in gateway, and your credit balance keeps whatever is in it. Switch between them in settings whenever you like.

What happens when subagents run at the same time?

Requests on one key share the stream's queue, one generation at a time. Parallel subtasks wait their turn instead of erroring or costing extra. Add streams if you want true concurrency.

Does this work in the JetBrains plugin and the Kilo CLI?

Yes. The provider is plain configuration, not an extension feature. Anywhere Kilo Code accepts an OpenAI-compatible base URL and key, Stream behaves the same.

What does a task cost on Stream?

Nothing beyond the flat rate. Every task, in every mode, is covered by the $5 a month you pay per stream. Any per-token estimate you see in the interface is based on public prices that don't apply here.

CAMELSTREAM · AGENT GUIDE

Load Kilo Code with unlimited frontier intelligence.

Kilo Code takes custom OpenAI-compatible providers right next to its built-in gateway. Add camelStream as one and every mode, in every editor you run Kilo in, works on frontier models for $5 a month per stream.

Get your Stream API key Explore camelStream

Setup

One custom provider.

Copy a Stream API key

Sign up for Stream, subscribe, and copy the key from your account. It's shown once, so save it before you switch back to the editor.

Add the custom provider

Open Kilo Code's settings, go to the Providers tab, and choose Custom provider at the bottom of the list. Fill in the provider ID, display name, base URL, and API key shown here.

Select auto from the picker

Once the URL and key are valid, Kilo Code fetches Stream's model list and offers auto. Select it and run a small task in a familiar repository before handing over a big one.

Kilo Code OpenAI-compatible provider docs

Kilo Code settings

Settings → Providers → Custom provider

# Provider ID:   camel-stream
# Display name:  camelStream
# Provider API:  OpenAI Compatible
# Base URL:      https://stream.camelai.com/v1
# API key:       <your Stream API key>

# Models: Kilo fetches Stream's model list
# and shows one entry, auto. Select it.

Base URL

stream.camelai.com/v1

Model ID

auto

Protocol

Chat Completions

Auth

Stream API key

Why Kilo Code on Stream

Stop topping up.

Kilo Code's own gateway is metered. The pricing is fair, but a balance that drains is still a balance you watch, and agent work drains it fastest. A planning pass hands work to subagents, each subtask carries its own context, and a busy afternoon shows up in your top-up history. Stream turns all of that into one flat line. Your credits stay for whatever you still want them for.

A balance that stops draining

Metered credits make you ration the agent without noticing you're doing it. At $5 a month per stream, the tenth task costs what the first one did.

Subtasks included

When a mode delegates to subagents, every subtask brings its own context and its own calls. On Stream the whole tree lands on the same flat price as a single prompt.

Every editor, same setup

Kilo Code runs in VS Code, JetBrains IDEs, and a CLI. The base URL and key on this page work in all of them, so where you code stops affecting what the model costs.

The deal

What every stream includes.

Frontier models only

You never pick a model; keeping the fleet smart is our job. Every model we serve scores 70% or higher on Terminal-Bench 2.1, or 50 or higher on the Artificial Analysis Intelligence Index, and always at its newest public release.

260K context, guaranteed

Every request gets at least a 260K-token window, and more when the model serving you supports it. Beyond the guaranteed window we compact the middle of the conversation first; your original task and latest turns stay intact.

One generation per stream

Unlimited works because each stream runs one generation at a time. Extra requests wait in a visible queue instead of raising your bill; add streams for parallel capacity.

$5 a month, flat

No token metering, no overage charges, cancel anytime. Sourcing efficiently across the fleet is what keeps the price flat.

The fleet changes as new models launch. The floor doesn't. The full spec sheet lives on the Stream page, and speed numbers are published targets, not a service-level agreement — the fine print is in the terms.

Troubleshooting

Check these first.

Most connection issues come down to the endpoint, API mode, model ID, or key. Start here before anything else.

Use https://stream.camelai.com/v1 as the base URL, ending at /v1 with nothing after it.

Keep the Provider API set to OpenAI Compatible. Stream serves Kilo Code over chat completions.

An empty model picker usually means the key is wrong. Kilo Code only fetches the model list once the base URL and key are both valid.

The model ID is auto, exactly as the picker shows it. Stream chooses the frontier model behind it.

You don't need kilo.jsonc overrides for Stream. Context and limits come from the model list itself.

Frequently asked questions

Kilo Code, answered.

More guides

Point another agent at Stream.

Claude Code

Keep the workflow. Change the bill.

Read the guide Codex

Native transport, separate profile.

Read the guide Aider

The meter was the whole problem.

Read the guide

Give Kilo Code the whole stream.

Unlimited frontier intelligence over the API format Kilo Code already speaks, for $5 a month per stream.

Get your API key