#1
Best AI agents in Research
Top Research agents independently tested by Hlido, ranked by overall score.
Quick answer: The top-scoring ai agents in research Hlido has independently tested is Exa (90/100, VITAL), ahead of Elicit (90/100). 20 agents ranked by hands-on test evidence, updated 2026-10-03.
#2
#3
wanshuiyin/Auto-claude-code-research-in-sleep
82/100
STEADY
Research
wanshuiyin/Auto-claude-code-research-in-sleep — ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works with Claude Code, Codex, OpenClaw, or any LLM agent.. Repo-surface review: measured from the public repository (13480 stars, MIT, last push 2026-07-14); not hands-on tested.
Pricing: Open source
#4
juanjuandog/FinSight-AI
82/100
STEADY
Research
juanjuandog/FinSight-AI — AI equity research agent with resilient workflows, Redis Lua single-flight, pgvector RAG, versioned reports, evidence tracing, and RAG evaluation.. Repo-surface review: measured from the public repository (1123 stars, MIT, last push 2026-05-26); not hands-on tested.
Pricing: Open source
#5
genieincodebottle/generative-ai
80/100
STEADY
Research
Public-surface review of genieincodebottle/generative-ai.
Pricing: Open source
#6
#7
#8
#9
#10
#11
#12
#13
Utopai-Research/pai-pro
67/100
FADING
Research
Utopai-Research/pai-pro — Local AI filmmaking studio — skills, canvas, timeline — driven from your coding agent.. Repo-surface review: measured from the public repository (319 stars, NOASSERTION, last push 2026-07-14); not hands-on tested.
#14
#15
#16
#17
Chronulus AI (MCP)
62/100
FADING
Research
Forecasting-and-prediction agents exposed to Claude via an MCP server, with a genuinely useful cold-start (zero-history) angle — but 'predict anything' is a broad claim the public surface cannot substantiate, and traction is still early.
Pricing: Paid
#18
Pensieve
60/100
FADING
Research
Pensieve should get a cautious T1 public-surface pass focused on visible claims, access friction, pricing clarity, and proof depth.
#19
Zaturn
58/100
FADING
Research
Open-source, local-first 'chat with your data' co-pilot that gives AI models SQL tools across many data sources, usable as an MCP server or a Jupyter-like studio — a promising, honest single-maintainer project that's early and asks you to bring your own LLM.
#20
Perplexity
53/100
FADING
Research
AI-powered answer engine with cited sources. Public no-login basic queries. Pro tier for advanced reasoning models.
Pricing: Free tier · Subscription
Why trust Hlido
Every score is derived from a fixed 5-dimension framework with C2PA-signed evidence captured during testing. We don't accept payment for placement.