docs: add strands-vllm community model provider documentation #448
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add documentation for strands-vllm, a vLLM model provider for Strands Agents SDK with Token-In/Token-Out (TITO) support for agentic RL training. Features documented: - OpenAI-compatible API integration with vLLM - TITO support for capturing prompt_token_ids and token_ids - Tool call validation hooks for RL-friendly error feedback - Agent Lightning integration via OpenTelemetry spans - TokenManager for building RL trajectories with loss masks References: - strands-vllm: https://github.com/agents-community/strands-vllm - Agent Lightning: https://github.com/microsoft/agent-lightning Closes strands-agents#432
zastrowm
reviewed
Jan 20, 2026
|
|
||
| The hook validates: | ||
|
|
||
| - **Tool name**: Must exist in agent's tool registry |
Member
There was a problem hiding this comment.
Does strands not do both of these automatically?
Specifically I think we do check for tool-names; do we not handle JSONDecodeError correctly - ore are you doing more here?
| agent = Agent(model=model, tools=[...], hooks=[VLLMToolValidationHooks()]) | ||
| ``` | ||
|
|
||
| ## Why TITO Matters |
Member
There was a problem hiding this comment.
Does this make sense here? (I'm not sure TBH).
It seems like if someone made it this far in the docs, they'd already know this - which makes me wonder if this should be moved up to the intro (but more concise) or removed from this doc.
Thoughts?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add documentation for strands-vllm, a vLLM model provider for Strands Agents SDK with Token-In/Token-Out (TITO) support for agentic RL training.
vllm.mdwith installation, usage, and configurationmkdocs.ymlnavigationCredits:
https://github.com/horizon-rl/strands-sglang/
Description
This PR adds community documentation for strands-vllm, a vLLM model provider designed for agentic reinforcement learning training workflows.
Key features documented:
/v1/chat/completionsendpoint with streamingprompt_token_idsandtoken_idsdirectly from vLLM - eliminates retokenization drift in RL trainingDocumentation includes:
VLLMToolValidationHooksVLLMTokenRecorderTokenManagerfor building trajectories with loss masksReferences:
Related Issues
Resolves #432
Related to #418 (strands-sglang - similar TITO provider pattern)
Type of Change
Checklist
mkdocs serveBy submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.