Model Catalog Migration 6/7: Agent Frameworks

[roh26it]

Model Catalog Migration - Agent Frameworks

What's changing

Agent framework integrations (CrewAI, Strands, OpenAI Agents, etc.) updated.

Files changed (10)

Major rewrites:

File	Changes
openai-agents.mdx	+284/-997
openai-agents-ts.mdx	+244/-741
crewai.mdx	+145/-632
strands.mdx	+150/-506

Medium changes:

• llama-agents.mdx, langchain-agents.mdx
• bring-your-own-agents.mdx, control-flow.mdx, phidata.mdx

Deleted:

• strands-backup.mdx (cleanup)

Review checklist

• Agent initialization uses Portkey client correctly
• Tracing/observability setup is documented
• Multi-provider configs use new format
• Handoff patterns documented correctly

Risk: 🟡 Medium

Growing adoption area - important to get right.

Part of

Model Catalog Migration - Split from #597 for easier review

[siddharthsambharia-portkey]

Hey @roh26it - for this PR, I suggest we standardize the index structure for each agent framework. Something like:

Proposed structure

1. Introduction

2. Quickstart

   • Model interoperability

3. Configuration

   • How to add configs / Portkey-specific headers for agents

4. Advanced Features
   (No agent-specific code here — this section should be reusable across all integrations via redirects or shared content.)

   • Observability (we can reuse the existing GIFs)
   • Prompt Management
   • Reliability features (fallbacks, load balancing, conditional routing, etc.)
   • Guardrails
   • Metadata
   • Caching (simple use cases only)

5. Enterprise Governance

   • Link out to a dedicated, framework-agnostic page

This keeps each integration lightweight, avoids duplication, and lets us copy-paste the same advanced/enterprise sections across frameworks with minimal maintenance.

Let me know if this structure works or if you want to tweak the grouping.