Metadata-Version: 2.4
Name: artest-llm
Version: 0.2.0
Summary: Provider gateway for language models: adapters, routing, call cache.
License-Expression: Apache-2.0
License-File: LICENSE
Requires-Dist: httpx>=0.28
Requires-Dist: opentelemetry-api>=1.45
Requires-Dist: pydantic>=2.14
Requires-Python: >=3.12
Project-URL: Repository, https://gitlab.com/Arta_Shoja/artest
Description-Content-Type: text/markdown

# artest-llm

Provider gateway for language models: adapters, routing, call cache.

It knows nothing about CRO. Agents declare what they need, the router picks a model, and the call cache
means a repeated call isn't paid for twice and can be replayed. A scheduler to keep every provider within its limits comes in
M3 ([#36](https://gitlab.com/Arta_Shoja/artest/-/issues/36)).

Import as `artest.llm`. See the [decision record](https://gitlab.com/Arta_Shoja/artest/-/blob/main/docs/adr/0005-llm-gateway.md).
