> ## Documentation Index
> Fetch the complete documentation index at: https://docs.celerflow.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# BioMed research pack

> PubMed-grounded research workflows: search, triage, PDF retrieval, and Zotero export.

The BioMed research pack is a dedicated set of seven [platform agents](https://docs.celerflow.ai/concepts/platform-agents.md) for life-science teams. It takes a biomedical question all the way from a PubMed search strategy to a Zotero-ready library of triaged papers with PDFs attached — including paywalled PDFs captured through your own browser session.

> [!NOTE]
>
> The BioMed pack is gated behind a paid plan. It is **not available on Free**; it is included on **Pro, Team, and Enterprise**. See [Billing](https://docs.celerflow.ai/platform/billing.md).

## When the Planner uses it

You don't pick the pack manually. When your request is grounded in PubMed / clinical literature — "find recent RCTs on…", "build a reading list on…" — the Planner reaches for these agents instead of the general-purpose web pipeline.

## The seven agents

| Agent | Role | What it does |
| - | - | - |
| `biomed_query_planner_agent` | analyzer | Turns a research topic into a reviewable PubMed search strategy. |
| `pubmed_search_agent` | collector | Searches PubMed via NCBI E-utilities and returns authoritative paper metadata. |
| `paper_triage_agent` | analyzer | Ranks candidate papers against your topic and explains why each one matters. |
| `citation_metadata_resolver_agent` | collector | Backfills missing citation metadata (DOI, PMID, PMC) for raw citations. |
| `pdf_access_resolver_agent` | collector | Determines whether a paper has an open-access / PMC PDF, or needs browser capture. |
| `institutional_access_agent` | collector | Creates a browser handoff session for papers behind a paywall or institutional login. |
| `zotero_library_agent` | sink | Prepares a Zotero RDF import bundle with the selected papers and PDF attachments. |

## The typical pipeline

```
biomed_query_planner → pubmed_search → paper_triage
  → pdf_access_resolver → [institutional_access for paywalled]
  → zotero_library
```

1. **Plan the search** — your topic becomes a structured, reviewable PubMed query.
2. **Search & triage** — PubMed returns candidates; triage ranks and explains them.
3. **Resolve PDFs** — open-access and PMC PDFs are fetched directly; paywalled ones are routed to a browser handoff.
4. **Capture paywalled PDFs** — you grab them in your own authenticated browser via the extension; the run resumes automatically.
5. **Export to Zotero** — the result is a Zotero RDF bundle you can import, papers and attachments included.

## Browser handoff

Step 4 is the only place a live run pauses for a human. CelerFlow does not bypass publisher access controls — paywalled PDFs are captured in *your* session, on your authority, through the [browser extension](https://docs.celerflow.ai/dashboard/browser-extension.md). See [Approvals](https://docs.celerflow.ai/concepts/approvals.md) for how the handoff pauses and resumes a run.

## See also

* [Platform agents](https://docs.celerflow.ai/concepts/platform-agents.md) — the full capability pool, including the general-purpose pack.
* [Browser extension](https://docs.celerflow.ai/dashboard/browser-extension.md) — installing and using the handoff extension.
* [Billing](https://docs.celerflow.ai/platform/billing.md) — which plans include the BioMed pack.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.