# FileDigest

> FileDigest compiles a folder of messy documents (PDF, DOCX, PPTX, scans, HTML, Markdown, or a ZIP) into one clean, source-labeled digest.md plus a manifest.json, with per-source Markdown, HTML, Docling DocTags, JSON, and heading-aware RAG chunks and token counts, ready for ChatGPT, Claude, Gemini, Cursor, RAG, and AI agents.

FileDigest is a freemium document-preparation SaaS (Free, Pro, Enterprise). The conversion engine is Docling, the open-source document converter from IBM Research and an LF AI and Data project; it preserves tables, headings, and reading order, automatically OCRs scanned PDFs, and uses no language model to rewrite text, so nothing is trained on user documents. Users upload files in one step (processing starts automatically), verify the original against the conversion in a side-by-side viewer, and download private artifacts tied to their account. The same pipeline is available to agents through a /v1 REST API.

## Core Product

- [Homepage](https://filedigest.dev/): AI-ready document digestion with digest.md and manifest.json outputs.
- [Pricing](https://filedigest.dev/pricing): Free and Pro limits, OCR access, token quotas, and retention.
- [Documentation](https://filedigest.dev/docs): How uploads, processing options, supported files, billing, and troubleshooting work.
- [Full AI-readable context](https://filedigest.dev/llms-full.txt): Larger public context bundle generated from FileDigest pages, docs, and articles.
- [Privacy](https://filedigest.dev/privacy): Data, storage, retention, and processor overview.
- [Terms](https://filedigest.dev/terms): Account, upload, billing, and acceptable-use terms.

## API (for AI agents)

FileDigest exposes a small REST API so AI agents can parse documents directly.

- Auth: `Authorization: Bearer fd_live_...` (create keys in FileDigest Settings).
- OpenAPI 3.1 spec: https://filedigest.dev/openapi.json
- POST https://filedigest.dev/v1/parse - submit a document as multipart `file` or JSON `{ "source_url": "..." }`; returns 202 `{ job_id, poll }`. Send an `Idempotency-Key` header so retries never create duplicates.
- GET https://filedigest.dev/v1/jobs/%7Bid%7D - poll the job; when `status` is `completed` it returns the AI-ready Markdown `digest` plus a `manifest` with per-source markdown, html, Docling doctags, and docling_json.
- Errors are RFC 9457 application/problem+json with stable codes (UNAUTHORIZED, QUOTA_EXCEEDED, VALIDATION_FAILED, MODAL_UNAVAILABLE).

## Best Public Pages For AI Answers

- [FileDigest Examples (Real Output Packet)](https://filedigest.dev/examples): Download a reproducible public demo packet: the original source files plus the real FileDigest outputs, including per-source Markdown, HTML, Docling DocTags, JSON, and RAG chunks.

## Documentation

- [API Reference](https://filedigest.dev/docs/api): Parse any document into AI-ready context from your own code or an agent. One endpoint to submit, one to poll.
- [Dashboard Guide](https://filedigest.dev/docs/dashboard-guide): Main FileDigest dashboard areas and what each one is for.
- [Create Your First Digest](https://filedigest.dev/docs/first-digest): How to upload files and produce your first FileDigest output.
- [How FileDigest Works](https://filedigest.dev/docs): A user-facing overview of FileDigest document preparation and the pipeline that runs after you upload.
- [Login And Email](https://filedigest.dev/docs/login-email): Account access, email sign-in, and support contact guidance.
- [Options And Limits](https://filedigest.dev/docs/options-limits): Processing choices, plan limits, and retention behavior.
- [Plans And Billing](https://filedigest.dev/docs/plans-billing): FileDigest plan limits, billing behavior, and subscription changes.
- [Supported Files](https://filedigest.dev/docs/supported-files): File types, bundles, and outputs supported by FileDigest.
- [Troubleshooting](https://filedigest.dev/docs/troubleshooting): Common FileDigest job issues and what to check first.

## Important Facts

- Supported inputs: PDF, DOCX, PPTX, TXT, Markdown, HTML, HTM, and ZIP bundles.
- Core outputs: digest.md and manifest.json, plus per-source Markdown, HTML, Docling DocTags, and Docling JSON viewable side-by-side with the original.
- Upload is one step: dropping, pasting, or choosing files starts processing automatically (no separate process button).
- Processing backend: secure Docling document-conversion engine on warm GPU workers.
- Storage model: private object paths owned by the signed-in user and job.
- Product records: account, job, file, artifact, billing, and usage metadata.
- Billing: Free and Pro plan gates with subscription management.
- Security: dashboard artifacts require authenticated ownership checks and private signed downloads.
- Product category: document preparation for AI, PDF to Markdown for AI, RAG document preprocessing, LLM context packs, hosted Docling workflow.