Science Agent Recommend: Find the Right AI Tool for Your Research
> Describe your research domain or task. Get personalized recommendations from 131+ curated AI science agents.
Decision Tree
Guide the user through these questions to narrow down recommendations:
Step 1: What's your research domain?
| Domain | Go to |
|---|
| -------- | ------- |
| Bioinformatics / Genomics / Omics | → Biomedicine & Omics |
| Drug Discovery / Chemistry / Molecular | → Drug & Molecular |
| General research / Literature / Writing | → Research & Paper Tools |
| Clinical / Medical / Health | → Clinical & Health |
| Education / Teaching | → Education |
| Evaluating / Benchmarking agents | → Benchmarks |
| Security / Auditing | → Security |
Step 2: Domain-Specific Recommendations
Biomedicine & Omics
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| General bioinformatics assistant | BioClaw | 318 | OpenClaw-based |
| Single-cell / multi-omics analysis | OmicsClaw | 114 | Python |
| Spatial transcriptomics | SpatialAgent | 158 | Genentech |
| CRISPR experiment design | CRISPR-GPT | 156 | Python |
| Protein structure + bio research | STELLA | 122 | Princeton |
| Nucleome / genome-scale | BioMaster | 91 | Python |
| Biomedical multi-tool (67 tools) | BioMedAgent | 58 | Nature BME paper |
Drug Discovery & Molecular
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| Drug-target interaction / pharmacology | DrugClaw (QSong) | 118 | 57 skills, LangGraph |
| Cheminformatics + folding + docking | BloClaw | 1 | All-in-one workspace |
| Pharma pipeline end-to-end | AIAgents4Pharma | 76 | Python |
| Molecular dynamics | MDCrow | 234 | OpenMM-based |
| Molecular visualization | PyMolClaw | 1 | 13 PyMOL scripts |
| Chemistry AI | ChemClaw | 32 | Python |
| TCM network pharmacology | TCM-Agent | 10 | PubChem + TTD |
Research & Paper Tools
Clinical & Health
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| Personal health copilot | HealthClaw | 293 | Streamlit + Feishu |
| Medical AI assistant | MedgeClaw | 953 | Python |
| Lab workflow automation | LabClaw | 935 | Python |
Education
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| Math learning (middle/high school) | MathClaw | 64 | Python, multi-channel |
| AI classroom | OpenMAIC | 14136 | Tsinghua |
Benchmarks & Evaluation
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| General agent evaluation | Claw-Eval | 339 | 300 tasks, 2159 rubrics |
| Agent safety (prompt injection) | ClawSafety Benchmark | 2 | 120 test cases |
| Bioinformatics agent evaluation | BioAgent Bench | 16 | arXiv paper |
| ML agent benchmark | HeurekaBench | — | EPFL |
Security
| Need | Recommended | Stars | Install |
|---|
| ------ | ------------ | ------- | --------- |
| Scan skills for vulnerabilities | ClawSafety Scanner | — | Rust CLI |
| Agent security monitoring | ClawKeeper | — | Security |
Step 3: Skill Hubs
If the user needs installable skills rather than full agents:
Browse all 32 skill hubs at claw4science.org/skill-hubs.
How To Respond
- Ask domain if not clear: "What's your research area?"
- Match to table: Find 2-3 best fits from the domain section
- Explain fit: One sentence on why each is relevant
- Show stars + link: Include Claw4Science listing URL
- Suggest skills: If they need specific skills, point to skill hubs
- Close with directory: "Browse all 131+ projects at claw4science.org"
API Access
# Get all projects with metadata
curl -s "https://claw4science.org/api/projects"
# Response includes: projects[], skills[], skill_hubs[]
# Each project has: title, description, repo, group, tags, static_stars
Links