# Best multi-agent frameworks for research agents

Research agents search, read, cross-check and write reports, usually by splitting a question across workers and merging the results. This page lists frameworks and harnesses whose records document that pattern. The order follows the criteria below. Data as of 2026-09-30.

## Selection criteria

- Documents a way to split a task across agents and merge results (supervisor, handoff, workforce or graph).
- Can persist or resume long runs.
- Connects to external tools and data (MCP evidence on the tool page).
- Runs on your own infrastructure with an open-source license.
- Supports human review of plans or tool calls.

## Shortlist

1. [DeerFlow](https://multiagentguide.top/tools/deer-flow.md): MIT-licensed harness built on LangGraph for research, reports and coding: a lead agent spawns sub-agents with their own context, `batch_task` runs large item sets from a durable SQL-backed queue, and MCP servers are configurable. Needs about 4 vCPU and 8 GB RAM per its README.
2. [LangGraph](https://multiagentguide.top/tools/langgraph.md): Graph runtime with checkpoints at every step, `interrupt()` for human review and time travel, which suits long research runs that must resume; multi-agent patterns are composed from subgraphs. It is low-level by its own description.
3. [Deep Agents](https://multiagentguide.top/tools/deepagents.md): LangChain's agent harness on LangGraph with subagents in isolated context, a pluggable filesystem and tool-call approval, so a long-horizon research agent does not have to be assembled from parts.
4. [LlamaIndex](https://multiagentguide.top/tools/llamaindex.md): `AgentWorkflow` handoffs on event-driven Workflows plus a large catalog of data integrations (the README counts over 300 packages) and a serializable run `Context`. The README notes the company's focus has moved to LlamaParse.
5. [CAMEL](https://multiagentguide.top/tools/camel.md): Apache-2.0 research framework whose Workforce splits tasks across worker agents; the official example is a searcher, analyst and writer team. A2A and AG-UI are unknown.
6. [Haystack](https://multiagentguide.top/tools/haystack.md): Explicit retrieval pipelines with an `Agent` component and a coordinator-plus-specialists pattern, breakpoints with JSON snapshots, and tool-level human confirmation.
7. [CrewAI](https://multiagentguide.top/tools/crewai.md): Role-based crews run sequentially or under a manager, with checkpointing and `@human_feedback` in Flows. Python >=3.10,<3.14.

## Shortlist facts

| Tool | Type | Languages | License | Pricing | Pattern | MCP | A2A | AG-UI | Stars |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| [DeerFlow](https://multiagentguide.top/tools/deer-flow.md) | Harness | Python, TypeScript | MIT | Open source, free | Supervisor | Yes | Unknown | Unknown | 83,261 |
| [LangGraph](https://multiagentguide.top/tools/langgraph.md) | Framework | Python | MIT | Open core | Graph | Yes | Yes | Partial | 42,511 |
| [Deep Agents](https://multiagentguide.top/tools/deepagents.md) | Framework | Python, TypeScript | MIT | Open core | Supervisor | Yes | Partial | Partial | 29,870 |
| [LlamaIndex](https://multiagentguide.top/tools/llamaindex.md) | Framework | Python | MIT | Open core | Handoff | Yes | Unknown | Yes | 52,370 |
| [CAMEL](https://multiagentguide.top/tools/camel.md) | Framework | Python | Apache-2.0 | Open source, free | Supervisor | Yes | Unknown | Unknown | 17,800 |
| [Haystack](https://multiagentguide.top/tools/haystack.md) | Framework | Python | Apache-2.0 | Open core | Supervisor | Yes | Unknown | Unknown | 26,631 |
| [CrewAI](https://multiagentguide.top/tools/crewai.md) | Framework | Python | MIT | Open core | Crew / roles | Yes | Yes | Partial | 59,217 |

## FAQ

### What is the difference between a research harness and a framework?

DeerFlow and Deep Agents ship a ready agent with sub-agents, sandbox or filesystem and memory. LangGraph, LlamaIndex, CAMEL, Haystack and CrewAI are libraries from which you build the research flow yourself.

### Which of these can run fully on my own machine?

All seven are open-source libraries or self-hosted apps. Model access is separate: each record lists whether local models (for example through Ollama) are documented.

### Do they support MCP for search and data tools?

Yes. Every tool on this list has an MCP cell marked yes with an evidence link on its page.

---

Data as of 2026-09-30. Not affiliated with listed projects. HTML version: https://multiagentguide.top/best/research-agents