CAMELSTREAM · AGENT GUIDE
Pair program on unlimited frontier intelligence.
Aider connects to any OpenAI-compatible endpoint with two environment variables. Set them to camelStream and every diff is drafted by a frontier model for $5 a month per stream. The cost column stops mattering.
Setup
Two variables and a flag.
Export your Stream credentials
Set CAMEL_API_KEY securely, then map it to the OPENAI_API_KEY variable Aider expects for an OpenAI-compatible provider.
Set the custom API base
Point OPENAI_API_BASE to https://stream.camelai.com/v1 before starting Aider in your repository.
Launch and make a small edit
Start Aider with openai/auto. Ask for a small, reviewable change first and inspect the diff before handing it a larger task.
export OPENAI_API_BASE="https://stream.camelai.com/v1"
export OPENAI_API_KEY="$CAMEL_API_KEY"
cd /path/to/your/repository
aider --model openai/autostream.camelai.com/v1
openai/auto
Chat Completions
env OPENAI_API_KEY
Why Aider on Stream
The meter was the whole problem.
Aider made a habit of showing you what each session costs, because on a per-token API you need to know. Point it at Stream and that number stops being interesting: map the repo, iterate on diffs, undo and retry freely. Your git workflow doesn't change at all.
Iterate on diffs, not budgets
Ask for the change, read the diff, ask again. Refining a change five times costs exactly the same as nailing it once.
Two variables, one flag
OPENAI_API_BASE, OPENAI_API_KEY, and a model name. It's the smallest integration on this site, and it leaves your git workflow untouched.
Repo maps welcome
Aider's repository map makes prompts long on purpose. With a 260K-token window guaranteed on every request, the map plus your files fit with room to spare.
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 OPENAI_API_BASE, not a URL that already includes /chat/completions.
Prefix the model with openai/ when launching Aider.
If Aider warns about an unfamiliar model, confirm the model ID and review Aider's model-warning guidance.
Keep your API key outside shell history and committed configuration files.
Frequently asked questions
Aider, answered.
More guides
Point another agent at Stream.
Give Aider the whole stream.
Unlimited frontier intelligence over the API format Aider already speaks, for $5 a month per stream.
Get your API key