Agent Loop docs
Installing
based-models-agentloop requires Python 3.12 or newer. Install the distribution named
based-models-agentloop and import it as agentloop.
For OpenAI, install the openai-compat extra:
python -m pip install 'based-models-agentloop[openai-compat]'
Check the import in the same Python environment:
python -c 'import agentloop; print("agentloop ready")'
Output:
agentloop ready
Extras
| Extra | Adds |
|---|---|
anthropic | Anthropic SDK |
openai-compat | HTTP client for OpenAI, OpenRouter, and Modal/vLLM |
otel | OpenTelemetry SDK and OTLP HTTP exporter |
all | All three integrations above |
The core package does not require these integrations. If you use a provider or telemetry feature without its dependency, the error names the extra to install. See the getting-started installation guide for more install commands and local development setup.
Pinning
Pin a compatible range in your project. While the version is 0.x, a minor release can
change the public API. For example, to stay on the 0.5 series with OpenAI support, use
this requirement in your project's dependency file:
based-models-agentloop[openai-compat]>=0.5,<0.6
For repeatable deployments, lock the selected version and its dependencies. If your project
uses pip-tools, pip-compile --generate-hashes can produce a hashed requirements file for
pip install --require-hashes.
Verifying a release
The PyPI files page shows
the wheel, source distribution, checksums, and provenance for the current version. The
matching GitHub release
carries the wheel, source distribution, and SHA256SUMS.