Scientific Agent Skills alternatives (2026)
Quick answer: The top picks on this list are AIPOCH Open-Science, Wisp Science and ChatAcademia. They lead 11 matching tools, ranked by GitHub stars on each listing. Pricing is what each maker states. Read each listing before you choose. Sources checked November 15, 2026. No vendor paid for a position.
Ranked comparison (top 10)
Dated .
| # | Tool | Pricing | Best for | Open source | Runs locally | GitHub stars |
|---|---|---|---|---|---|---|
| 1 | Open source | control a computer, research the web | Yes | Offline | 5,469 | |
| 2 | Open source | research the web | Yes | Offline | 1,059 | |
| 3 | Freemium | research the web | No | |||
| 4 | Open source | write code, research the web | Yes | 4,303 | ||
| 5 | Open source | write code, build an agent | Yes | Self-hosted | 84,715 | |
| 6 | Open source | control a computer, research the web | Yes | Offline | 38,419 | |
| 7 | Open source | build an agent, research the web | Yes | 3,455 | ||
| 8 | Open source + cloud | build an agent, research the web | Yes | Self-hosted | 4,192 | |
| 9 | Open source | build an agent, research the web | Yes | Self-hosted | 3,124 | |
| 10 | Paid | research the web | No |
How we rank: GitHub stars on each listing. Stars and maturity come from each AgentScout listing. No vendor paid for a position.
Updated / sources checked .
Scientific Agent Skills is a research tool: Scientific Agent Skills, formerly Claude Scientific Skills, is K-Dense's MIT-licensed library of 177 validated Agent Skills plus connectors to more than 100 scientific databases for biology, chemistry, medicine and drug discovery; it works with Cursor, Claude Code, Codex and any agent that supports the open Agent Skills standard, and also powers K-Dense BYOK, a free desktop co-scientist that runs on your own API keys. Pricing, as the maker states it: Open source. Its GitHub repo has 56,830 stars. We found 62 tools in the directory that do overlapping work (controlling a computer, building AI agents, research); 45 of them are open source and 12 list a free or freemium plan. Below are the 10 closest, ranked by how many tasks and sub-tasks they share with Scientific Agent Skills, then by GitHub stars.
Updated
AIPOCH Open-Science (Open source): A local-first, model-agnostic desktop research workbench. You work with an agent on papers and data through skills, MCP tools, and Python/R execution, then export traceable .science packages for reproducible results. It runs on macOS, Windows and Linux. Consider it because it also covers research and reports.
Wisp Science (Open source): Wisp Science is an open-source, local-first AI research workbench for scientists on Windows, macOS and Linux. Its agent searches the literature, runs persistent Python and R kernels on your laptop or remote servers, queries about 80 scientific databases and loads reusable skills, while keeping figures, runs, decisions and drafts together in one project. You bring your own OpenAI-compatible or Anthropic model, or drive Codex or Claude Code, and approval gates stay on by default. Consider it because it is ready to use without building anything; it also covers research and reports.
ChatAcademia (Freemium): ChatAcademia is a set of AI research agents for academics. Its agents search many scholarly databases, pull insights out of papers and help with each step from forming an idea to drafting a publication, and a desktop app can work directly with the research documents on your own computer. Consider it because it is ready to use without building anything; it also covers RAG and documents and research and reports.
OpenScience (Open source): Research agent workbench that reads the literature, writes and runs code and experiments and logs every step, with skills and connectors for databases like PubMed, UniProt and arXiv. Consider it because it is ready to use without building anything; it also covers research and reports.
Deer Flow (Open source): DeerFlow 2.0 is ByteDance's MIT-licensed super-agent harness for long tasks that can take minutes to hours, such as research reports, coding jobs and content creation. It coordinates sub-agents, long-term memory, sandboxed code execution, tools and extensible skills, and can take requests through a message gateway; version 2.0 is a ground-up rewrite, with the original Deep Research framework kept on a separate branch. Consider it because it also covers memory and research and reports; it has more GitHub stars (84,715 vs 56,830).
PDFMathTranslate (Open source): PDFMathTranslate (pdf2zh) translates scientific PDFs while keeping the original layout: formulas, charts, tables of contents and annotations stay in place and it outputs bilingual or translated-only files. It works with Google, DeepL, OpenAI, Ollama and other translation services, adds experimental OCR and an ultrafast mode for native-text PDFs, and runs as a CLI, a desktop or web GUI, Docker, an MCP server or a Zotero plugin. AGPL-3.0; an EMNLP 2025 demo paper. Consider it because it is ready to use without building anything; it also covers RAG and documents and MCP servers.
LLM for Zotero (Open source): Zotero plugin that adds a research agent to your library so you can ask questions about papers, compare sources and save notes without leaving the reader. Consider it because it is ready to use without building anything; it also covers research and reports.
Hyperresearch (Open source + cloud): Deep research harness that turns Claude Code or Codex into a research agent building a persistent, searchable wiki, with a hosted API and MCP. Consider it because it also covers research and reports.
MiroFlow (Open source): Research agent framework with web UI supporting MiroThinker and other models. Consider it because it also covers research and reports.
AskYourDatabase (Paid): Chat-with-your-database tool that answers natural-language questions over Postgres, MySQL, BigQuery, Snowflake and more with SQL and charts. Consider it because it is a managed commercial product, if you'd rather not run it yourself; it is ready to use without building anything.
Scientific Agent Skills vs alternatives at a glance
| Tool | Pricing | Open source / licence | GitHub stars | Good for |
|---|---|---|---|---|
| Open source | Yes | 56,830 | controlling a computer, building AI agents | |
| Open source | Yes | 5,469 | controlling a computer, research | |
| Open source | Yes | 1,059 | research, reading scientific papers | |
| Freemium | No | — | research, reading scientific papers | |
| Open source | Apache License 2.0 | 4,303 | writing code, research | |
| Open source | MIT | 84,715 | writing code, building AI agents | |
| Open source | Yes | 38,419 | coding in the terminal, controlling a computer | |
| Open source | Yes | 3,455 | building AI agents, research | |
| Open source + cloud | MIT | 4,192 | building AI agents, research | |
| Open source | Apache License 2.0 | 3,124 | building AI agents, research | |
| Paid | No | — | research, data analysis |
More
Related searches
Alternatives are picked from directory data only: same category plus shared tasks and sub-tasks. Pricing is as each maker states it; GitHub stars as recorded on each listing.