# SalesforceAIResearch/enterprise-deep-research

Salesforce Enterprise Deep Research

Repository: https://github.com/SalesforceAIResearch/enterprise-deep-research
Canonical: https://ross.abutalabs.com/products/enterprise-deep-research
Language: Python
License: Apache-2.0
License Family: permissive
Topics: deep-research-agent, llm-benchmarking, multi-agent-systems, e2b, fastapi, langchain, react, tailwindcss, tavily
Last push: 2026-06-02T18:57:29+00:00

## Health v2 (maintenance only)
Score: 56/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 85, release rhythm 35, longevity 25
- inputs: {"age_days": 351, "days_push": 92, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1203, forks 191 (observed 2026-08-28T04:03:58.225465+00:00)

## What it is
Enterprise Deep Research (EDR) is a multi-agent deep research system from Salesforce AI Research that combines a master planning agent, specialized search agents, MCP-based tools, and a visualization agent to produce automated research reports. It includes a FastAPI backend and React web UI, supports real-time steering and human-in-the-loop guidance, and ranks #1 on LiveResearchBench.

## Use cases
- generate comprehensive research reports on complex queries automatically
- run multi-agent web research with academic, GitHub, and LinkedIn search
- benchmark deep research agents on EDR-200 and leaderboards
- steer ongoing research in real time with human feedback
- analyze enterprise data with NL2SQL and file analysis tools
- self-host an AI research assistant for my organization

## When to choose
- you need automated, citation-backed deep research reports
- you want a self-hosted multi-agent research pipeline with extensible MCP tools
- you need human-in-the-loop steering during long research runs
- you are evaluating deep research systems against benchmarks

## When to avoid
- you need a simple single-shot LLM query without research orchestration
- you cannot provide LLM API keys or self-host a Python/React stack
- you need a lightweight chatbot rather than a full research workflow

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, rag, search-engine, llm-inference, data-visualization, benchmarking, mcp
- domain: artificial-intelligence, large-language-models, erp, developer-tools
- platform: python, self-hosted
- tags: deep-research, multi-agent-system, report-generation, human-in-the-loop, fastapi, langchain, web-search, nl2sql, enterprise-ai, ai-agents, search, web-server, docker

## Member repositories
- SalesforceAIResearch/enterprise-deep-research (main) score 56

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:58.225465+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:19:43.147431+00:00, confidence not recorded.
  - readme: https://github.com/SalesforceAIResearch/enterprise-deep-research (fetched 2026-08-28T04:03:58.225465+00:00, sha 7ac148c9efd2)
- Data as of 2026-08-30T08:39:29.467469+00:00.