[Integrations](https://spiral.ai/integrations) › [Codex](https://spiral.ai/integrations/codex) › [Firecrawl paper search](https://spiral.ai/integrations/codex/firecrawl-paper-search)

# How to Use Firecrawl paper search in Codex

Firecrawl paper search returns ranked papers with abstracts, authors, source IDs, dates, and relevance scores. Use it to find literature from PubMed, bioRxiv, medRxiv, and arXiv rather than ordinary web snippets.

|  |  |
| --- | --- |
| Model | Firecrawl paper search |
| Available through | Firecrawl |
| Category | Data |
| Price | $0.013 per call. |
| Returns | json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Firecrawl paper search in Codex

- Use Firecrawl paper search to find 15 papers on HNSW approximate nearest-neighbor search. Read our vector search design notes and add an abstract-based bibliography with titles, authors, and source IDs.
- Use Firecrawl paper search to find 12 papers about retrieval-augmented generation answer faithfulness. Compare their abstracts with our evaluation plan and suggest papers to read before changing the tests.
- Use Firecrawl paper search to find the paper Attention Is All You Need. Add its returned title, authors, abstract, and source identifier to our research notes.
- Use Firecrawl paper search to find 10 papers about LoRA and QLoRA. Review our fine-tuning proposal and create a reading list grouped by the methods described in the abstracts.
- Use Firecrawl paper search to find 15 papers about HumanEval code generation evaluation. Draft a bibliography for our benchmark documentation and identify candidate papers whose abstracts discuss evaluation limitations.

## How to set up Firecrawl paper search in Codex

Pick the way that suits you. Each takes about two minutes.

### Use Spiral's MCP server

1. **Add the hosted server**

```bash
codex mcp add spiral --url https://api.spiral.ai/mcp
```

2. **Sign in**: Codex opens the browser: sign in to Spiral and click Authorize. No API key. `codex mcp list` then shows Spiral.

```bash
codex mcp login spiral
```

### Use the Spiral skill

1. **Send Codex this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run Firecrawl paper search**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "firecrawl",
    "model": "firecrawl/research-search",
    "input": {
      "query": "<query>"
    }
  }'
```

Then paste one of the prompts above: Codex finds Firecrawl paper search on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/codex

## Firecrawl paper search tools and prices

- [Firecrawl Research · Paper Search](https://spiral.ai/tools/firecrawl/firecrawl/research-search) (Firecrawl): $0.013 per call.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| k | 1 to 500 | 40 | Maximum number of ranked papers to return. |
| to |  | — | Inclusive upper bound on created/updated date. |
| from |  | — | Inclusive lower bound on created/updated date. |
| sort | relevance, date_desc | relevance | Result order. `relevance` (default) returns the most relevant papers first. `date_desc` returns the newest papers among the most relevant matches, newest first by created date, with undated papers last. It does not list every paper in a date range; use `from` and `to` for that. |
| authors |  | — | Author substring filter. Repeat or pass a comma-separated value; all filters must match. |
| categories |  | — | Paper category filter. Repeat or pass a comma-separated value; all filters must match. |

## What can you do with Firecrawl paper search in Codex?

Build a research bibliography beside the code that needs it, using Firecrawl's paper titles, abstracts, and identifiers. A coding agent can connect candidate papers to an implementation proposal and explain which claims come from abstracts.

### Build a literature review shortlist

Find papers on retrieval-augmented generation evaluation, then organize their abstracts into retrieval quality, answer faithfulness, and benchmark design. Keep the returned paper identifiers beside each entry for later reading.

### Find papers behind an engineering decision

Search for approximate nearest-neighbor methods before choosing an approach for a vector search project. Build a bibliography of candidate methods and summarize the tradeoffs described in their abstracts.

### Prepare a journal club reading list

Find papers about CRISPR base-editing off-target effects and select abstracts that describe different experimental approaches. Return titles, authors, dates, and identifiers so the club can choose what to read.

### Recover a paper from its description

Search for a remembered method, such as diffusion models outperforming GANs on image synthesis. Compare candidate titles and abstracts to identify the likely paper.

### Find benchmark papers

Collect papers about MMLU or HumanEval and use their abstracts to separate benchmark introductions from papers applying those benchmarks. Produce a reading list for an evaluation project.

### Maintain a topic digest

Repeat searches for wastewater-based epidemiology within recent date windows and compare returned paper IDs with earlier results. Summarize newly found abstracts as a selected reading digest.

### Compare research approaches

Run separate searches for LoRA and full fine-tuning, then compare the methods and applications described in the returned abstracts. Use the comparison to choose papers for a deeper technical review.

## Overview

- [Firecrawl](https://spiral.ai/providers/firecrawl)
- [All Codex guides](https://spiral.ai/integrations/codex)

## Firecrawl paper search in other agents

- [Amp](https://spiral.ai/integrations/amp/firecrawl-paper-search)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/firecrawl-paper-search)
- [Claude](https://spiral.ai/integrations/claude-desktop/firecrawl-paper-search)
- [Claude Code](https://spiral.ai/integrations/claude-code/firecrawl-paper-search)
- [Cline](https://spiral.ai/integrations/cline/firecrawl-paper-search)
- [Cursor](https://spiral.ai/integrations/cursor/firecrawl-paper-search)
- [Devin](https://spiral.ai/integrations/devin/firecrawl-paper-search)
- [dots](https://spiral.ai/integrations/dots/firecrawl-paper-search)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/firecrawl-paper-search)
- [Gemini](https://spiral.ai/integrations/gemini/firecrawl-paper-search)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/firecrawl-paper-search)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/firecrawl-paper-search)
- [goose](https://spiral.ai/integrations/goose/firecrawl-paper-search)
- [Grok](https://spiral.ai/integrations/grok/firecrawl-paper-search)
- [Grok Build](https://spiral.ai/integrations/grok-build/firecrawl-paper-search)
- [Hermes](https://spiral.ai/integrations/hermes/firecrawl-paper-search)
- [Instinct](https://spiral.ai/integrations/instinct/firecrawl-paper-search)
- [Manus](https://spiral.ai/integrations/manus/firecrawl-paper-search)
- [Muse](https://spiral.ai/integrations/muse/firecrawl-paper-search)
- [Muse Code](https://spiral.ai/integrations/muse-code/firecrawl-paper-search)
- [n8n](https://spiral.ai/integrations/n8n/firecrawl-paper-search)
- [Notion](https://spiral.ai/integrations/notion/firecrawl-paper-search)
- [OpenClaw](https://spiral.ai/integrations/openclaw/firecrawl-paper-search)
- [opencode](https://spiral.ai/integrations/opencode/firecrawl-paper-search)
- [Perplexity](https://spiral.ai/integrations/perplexity/firecrawl-paper-search)
- [Raycast](https://spiral.ai/integrations/raycast/firecrawl-paper-search)
- [VS Code](https://spiral.ai/integrations/vscode/firecrawl-paper-search)
- [Windsurf](https://spiral.ai/integrations/windsurf/firecrawl-paper-search)
- [Zapier](https://spiral.ai/integrations/zapier/firecrawl-paper-search)

## More Codex guides

- [Firecrawl government documents](https://spiral.ai/integrations/codex/firecrawl-government-documents)
- [Firecrawl related research papers](https://spiral.ai/integrations/codex/firecrawl-related-research-papers)
- [Bria editable ad layers](https://spiral.ai/integrations/codex/bria-editable-ad-layers)
- [Firecrawl research paper metadata and passages](https://spiral.ai/integrations/codex/firecrawl-research-paper-metadata-and-passages)
- [Bria Fibo structured image prompt](https://spiral.ai/integrations/codex/bria-fibo-structured-image-prompt)
- [Gemini content classification](https://spiral.ai/integrations/codex/gemini-content-classification)

## Questions about Firecrawl paper search in Codex

**Can Codex use Firecrawl paper search?**

Yes. Add Spiral to Codex as an MCP server, and Codex can find and run Firecrawl paper search like any built-in tool.

**How much does Firecrawl paper search cost in Codex?**

$0.013 per call. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Codex plan is separate.

**Do I need a Firecrawl account to use Firecrawl paper search in Codex?**

No. Spiral runs Firecrawl paper search on its own Firecrawl account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Firecrawl paper search in Codex?**

About two minutes: add the hosted server, then sign in.

**Does Codex wait while Firecrawl paper search runs?**

Firecrawl paper search runs in the background. Codex starts the run, checks on it and gives you the json when it's ready.

**Which research sources does Firecrawl cover?**

The index contains roughly 43 million abstracts, around 90% biomedical and life sciences literature from PubMed, bioRxiv, and medRxiv. arXiv adds physics, mathematics, and computer science papers.

**What does a paper search result include?**

Results include the title, abstract, authors, source identifiers, creation and update dates, and a ranking score. These are ranked paper records rather than ordinary web snippets.

**How many papers can one search return?**

One search can return up to 500 ranked papers. That ceiling is a result limit, not a count of all literature matching the topic.

**Does newest-first ordering list every recent paper?**

No. Firecrawl orders the most relevant matches by creation date, so the result remains a relevance-selected list rather than a complete chronological feed.

**Why might a known paper be missing?**

Coverage outside PubMed, bioRxiv, medRxiv, and arXiv is thinner. Firecrawl specifically warns that papers available only behind publisher paywalls or in niche venues may not be indexed.

