# jina-ai/node-DeepResearch

Keep searching, reading webpages, reasoning until it finds the answer (or exceeding the token budget)

Repository: https://github.com/jina-ai/node-DeepResearch
Canonical: https://ross.abutalabs.com/products/node-deepresearch
Homepage: https://search.jina.ai
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: deepresearch, deepsearch
Last push: 2026-05-01T20:59:41+00:00

## Health v2 (maintenance only)
Score: 58/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 80, release rhythm 40, longevity 41
- inputs: {"age_days": 584, "days_push": 124, "days_rel": 567, "gap_med": 0.0, "n_releases_24m": 7}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5225, forks 461 (observed 2026-08-28T04:09:13.259562+00:00)

## What it is
A TypeScript agent that iteratively searches the web, reads webpages, and reasons with an LLM (Gemini, OpenAI, or local) until it finds a concise answer or exceeds its token budget. It powers Jina AI's hosted DeepSearch UI and API and is intended for quick, factual answers rather than long-form reports.

## Use cases
- build a deep search agent that keeps searching until it finds the answer
- self-host an alternative to OpenAI/Perplexity deep research
- get concise cited answers to complex queries via iterative web search
- integrate deep search into an app via a chat-completions-style API
- learn how DeepSearch/DeepResearch agents are implemented
- add a research agent to my browser search engines

## When to choose
- you need quick, concise, well-sourced answers rather than long AI-generated reports
- you want a small, hackable TypeScript codebase to study or extend agentic search loops
- you want to self-host or run locally with Gemini, OpenAI, or a local LLM

## When to avoid
- you need long-form multi-section research reports like OpenAI/Gemini Deep Research
- you want a polished production product without touching code - the hosted Jina API may fit better
- you need a Python-based agent stack

## Facets
- artifact type: application
- maturity: active
- function: agent-framework, search-engine, llm-inference, web-scraping, rag
- domain: artificial-intelligence, developer-tools
- platform: cross-platform
- tags: deep-research, deep-search, iterative-search, web-reading, gemini, openai, jina-reader, answer-engine, search, natural-language-processing, nodejs, typescript, web-server

## Member repositories
- jina-ai/node-DeepResearch (main) score 58

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:13.259562+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 2026-08-29T17:59:41.601778+00:00, confidence not recorded.
  - readme: https://github.com/jina-ai/node-DeepResearch (fetched 2026-08-28T04:09:13.259562+00:00, sha 6728200c0b66)
  - homepage: https://search.jina.ai (fetched 2026-08-29T08:55:15.429553+00:00, sha 7286d92d9538)
- Data as of 2026-08-30T08:39:29.467469+00:00.