CAMELSTREAM · AGENT GUIDE
Run Hermes Agent on unlimited frontier intelligence.
Hermes Agent can run its whole tool loop through one custom OpenAI-compatible endpoint. Point it at camelStream and that loop runs on frontier models for $5 a month per stream, no token meter attached.
Setup
Five fields in the wizard.
Create your Stream API key
Sign up for Stream, subscribe to the founding plan, and copy the API key from your Stream account. Keys are shown once, so store it somewhere safe.
Walk the hermes model wizard
Run hermes model, choose Custom endpoint, set the API mode to Chat Completions, and enter the base URL, model ID, context length, and API key shown here.
Start a tool-using session
Open Hermes in a test repository and ask it to inspect files before making a small change. Confirm that text generation and tool calls both complete.
hermes model
# Choose: Custom endpoint
# Base URL: https://stream.camelai.com/v1
# API mode: Chat Completions
# Model: auto
# Context length: 262144
# API key: <your Stream API key>stream.camelai.com/v1
auto
Chat Completions
Stream API key
Why Hermes Agent on Stream
Built for the long loop.
Hermes Agent is at its best when you let it work: reading files, running commands, retrying until the tests pass. That style burns tokens by design, and on a metered API every retry has a price on it. On Stream the loop costs the same whether it lands in one attempt or forty.
Let it retry
Inspect, run, fail, fix, run again. The agent's persistence stops being a line item, so you can leave ambitious tasks running without watching a spend dashboard.
One wizard, no rewrite
Hermes documents custom OpenAI-compatible endpoints in its model wizard. You change five fields; the rest of your agent setup stays exactly as it was.
A queue instead of a bill
Each stream runs one generation at a time and queues the rest. A burst of parallel experiments never turns into overage charges; add streams when you want real concurrency.
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 exactly as shown, including /v1 but not /chat/completions.
Set the API mode to Chat Completions. Other Hermes API modes will 404 against this endpoint.
Use auto as the model ID; model names copied from another provider will not resolve.
If Hermes was already running, restart it after changing the saved provider or API key.
Test the main model first, then verify any separately configured Hermes auxiliary models.
Frequently asked questions
Hermes Agent, answered.
More guides
Point another agent at Stream.
Give Hermes Agent the whole stream.
Unlimited frontier intelligence over the API format Hermes Agent already speaks, for $5 a month per stream.
Get your API key