mirror of
https://github.com/BerriAI/litellm.git
synced 2026-09-27 01:22:18 +00:00
Managed batch ownership rows were already written for every create path, but retrieve/cancel/file-content checks only fired for unified (litellm_proxy-prefixed) ids, so model-encoded and raw provider batch ids bypassed isolation entirely. - enforce can_access_resource on retrieve/cancel for provider-format batch ids when an ownership row exists; ids with no row stay accessible so pass-through reads keep working - make batch sync operations update-only so a retrieve/cancel can never mint an ownership row attributed to the first caller - write ownership rows for a synced batch's provider-format output/error file ids, inherited from the owning batch row, and enforce them on file content/retrieve/delete |
||
|---|---|---|
| .. | ||
| cloudformation_stack | ||
| dist | ||
| enterprise_hooks | ||
| enterprise_ui | ||
| litellm_enterprise | ||
| __init__.py | ||
| LICENSE.md | ||
| pyproject.toml | ||
| README.md | ||
LiteLLM Enterprise
Code in this folder is licensed under a commercial license. Please review the LICENSE file within the /enterprise folder
These features are covered under the LiteLLM Enterprise contract
👉 Using in an Enterprise / Need specific features ? Meet with us here
See all Enterprise Features here 👉 Docs