docs: add Entroly context optimization integration

[juyterman1000]

What this PR adds

Adds Entroly as an integration tool in the CrewAI docs.

Files changed:

• New: docs/en/tools/integration/entrolytool.mdx — Full integration guide
• Modified: docs/en/tools/integration/overview.mdx — Added card to the integration overview

Why this fits CrewAI:

Multi-agent CrewAI workflows multiply token costs because each agent independently sends large context windows. Entroly addresses this specifically:

• 70-95% fewer input tokens via local context compression
• Cache alignment — keeps prefixes stable so provider cache discounts apply (Anthropic 90%, OpenAI 50%)
• Nash-KKT budget allocation — optimally splits token budgets across multiple agents
• WITNESS hallucination guard — evidence-grounding check per response

Integration is transparent:

\\�ash
pip install entroly && entroly proxy
\\

\\python
os.environ['OPENAI_BASE_URL'] = 'http://localhost:9377/v1'

CrewAI works as normal — Entroly compresses context transparently

\\

Apache-2.0, local-first, no outbound analytics by default. All benchmark claims backed by committed JSON artifacts.

Summary by CodeRabbit

• Documentation
  • Added comprehensive documentation for Entroly Context Optimization integration with CrewAI, including setup instructions for proxy and library modes
  • Provides guidance on configuring the local proxy and using environment variables
  • Documents dashboard access and key features for local context compression to reduce LLM API costs

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR adds documentation for the Entroly Context Optimization integration with CrewAI. It introduces a new detailed documentation page explaining Entroly's context compression capabilities and provides setup instructions for both proxy and library modes, plus references the local dashboard and feature summary.

Changes

Entroly Integration Documentation

Layer / File(s)	Summary
Integration directory entry docs/en/tools/integration/overview.mdx	Entroly Context Optimization card added to the tools overview grid with icon, destination URL, and capability description.
Documentation page foundation docs/en/tools/integration/entrolytool.mdx	Frontmatter metadata, introduction to Entroly's context compression and multi-agent token budgeting mechanisms, and pip install entroly installation instructions.
Setup and configuration guides docs/en/tools/integration/entrolytool.mdx	Quick Setup section with proxy mode commands and CrewAI environment variable configuration; Library Mode example using compress_messages; local dashboard command documentation.
Features and resources docs/en/tools/integration/entrolytool.mdx	Key Features table summarizing context compression, cache alignment, WITNESS guard, multi-agent budget allocation, and local-first operation; Resources section with links to GitHub, documentation, and benchmarks.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Poem

🐰 A new tool joins the CrewAI crew,
Entroly compresses what tokens do,
From proxy to library, docs shine bright,
Context compression—done right!
Hop-hop, our integration's complete! 🌟

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'docs: add Entroly context optimization integration' clearly and accurately summarizes the main change: adding documentation for a new Entroly integration tool.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

Warning

⚠️ This pull request might be slop. It has been flagged by CodeRabbit slop detection and should be reviewed carefully.

[coderabbitai]

🧹 Nitpick comments (1)

docs/en/tools/integration/overview.mdx (1)

25-27: ⚡ Quick win

Consider qualifying the "70-95%" cost reduction claim.

The percentage range is specific and prominent, but the overview card doesn't indicate which scenarios achieve the higher vs lower end of the range. Readers might expect 70-95% savings across all use cases. Other integration cards in this overview focus more on capabilities ("access hundreds of tools", "invoke live automations") rather than quantified outcome claims.

While the PR objectives note this claim is backed by benchmarks, consider whether the overview card should either:

• Add brief context: "Reduce LLM API costs by up to 70-95% depending on workload..."
• Or soften to: "Significantly reduce LLM API costs with local context compression..."
• Or defer the percentage to the detailed page where benchmark context can be explained

This helps set realistic expectations before users click through to the full documentation.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/en/tools/integration/overview.mdx` around lines 25 - 27, The overview
Card with title "Entroly Context Optimization" makes a specific "Reduce LLM API
costs by 70-95%" claim without context; update the Card copy to either add
qualifying context ("Reduce LLM API costs by up to 70-95% depending on workload
and cache hit rates") or soften it to a non-quantified statement ("Significantly
reduce LLM API costs with local context compression"), or remove the percentage
here and move the benchmark details to the linked page
(/en/tools/integration/entrolytool) so the overview remains consistent with
other cards; edit the Card element text accordingly to reflect the chosen
approach.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@docs/en/tools/integration/overview.mdx`:
- Around line 25-27: The overview Card with title "Entroly Context Optimization"
makes a specific "Reduce LLM API costs by 70-95%" claim without context; update
the Card copy to either add qualifying context ("Reduce LLM API costs by up to
70-95% depending on workload and cache hit rates") or soften it to a
non-quantified statement ("Significantly reduce LLM API costs with local context
compression"), or remove the percentage here and move the benchmark details to
the linked page (/en/tools/integration/entrolytool) so the overview remains
consistent with other cards; edit the Card element text accordingly to reflect
the chosen approach.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 01d46df4-ae68-427a-8aa1-a9e6c7d62a28

📥 Commits

Reviewing files that changed from the base of the PR and between 906cd97 and 39af455.

📒 Files selected for processing (2)

• docs/en/tools/integration/entrolytool.mdx
• docs/en/tools/integration/overview.mdx