# Best Agent frameworks — Opt out of training

> Source: https://theagentsindex.com/best/frameworks/training (curated, quality-gated, re-verified)

7 of 16 agent frameworks have a confirmed yes for “Opt out of training”. 6 are qualified or tier-scoped, 0 are confirmed no, and 3 remain not established. Only confirmed yes answers appear in this set; an unknown is never treated as a negative.

This is not an arbitrary slice of agent frameworks. Among the ones with a confirmed yes for “Opt out of training”, 7 of 7 also answer yes on how it meters, against 1 of 9 across the rest of the category. Every figure is counted from the stored, sourced datasheet answers linked on each row.

| # | Name | What it is | Pricing | API |
| --- | --- | --- | --- | --- |
| 1 | [LangGraph](https://theagentsindex.com/langgraph.md) | Open-source orchestration framework and runtime for stateful, long-running LLM agents, from LangChain. | freemium | Yes |
| 2 | [Agno](https://theagentsindex.com/agno.md) | Python SDK, open-source AgentOS runtime and hosted control plane for building and operating agents. | freemium | Yes |
| 3 | [Claude Agent SDK](https://theagentsindex.com/claude-agent-sdk.md) | Python and TypeScript library that runs Claude Code's agent loop and tools inside your own application. | paid | Yes |
| 4 | [Mastra](https://theagentsindex.com/mastra.md) | Open-source TypeScript framework for building AI agents, with a hosted platform for running and observing them. | freemium | Yes |
| 5 | [LiveKit Agents](https://theagentsindex.com/livekit-agents.md) | Open-source Python and Node.js framework for voice AI agents, with LiveKit Cloud hosting, telephony and observability. | freemium | Yes |
| 6 | [LlamaIndex](https://theagentsindex.com/llamaindex.md) | Open-source agent framework plus a credit-metered document parsing, extraction and indexing platform. | freemium | Yes |
| 7 | [LangChain](https://theagentsindex.com/langchain.md) | Open source agent frameworks plus LangSmith, a hosted platform to trace, evaluate, deploy and govern agents. | freemium | Yes |

## Methodology

Projected from the live listing datasheets using the tenant’s “Can you stop your code training someone else’s model?” axis. Membership requires a sourced yes; partial and unknown answers are reported above but excluded from the ranking. Differences from the rest of the category are Fisher exact tests over 34 other datasheet axes, Holm-corrected. Entries retain the directory’s paid-tier ordering, visibly labelled on every promoted card.