Your documents. Your server. Your rules.

Syndoc runs on your infrastructure and never phones home. It doesn't just search your documents — it reasons across them with agentic AI, extracts the data, and runs the workflows around them. Licensed once, owned forever.

Used by teams in legal, finance, and healthcare administration.

syndoc — document chat

You

What's the total invoice amount across Q2 contracts?

Planning

Decomposed into 3 sub-queries · hybrid BM25 + vector · searched 8 documents

Syndoc

Across 3 Q2 contracts, invoiced totals come to $412,890.00.

Sources: 3 documents cited

  • You're paying monthly for tools that own your data.

  • Your legal team doesn't trust that AI chatbot you subscribed to.

  • You're stitching together four vendors to do one job.

There's a better way to do this.

What Syndoc is

One platform. Your infrastructure. Permanent license.

Syndoc runs on your servers — Azure, AWS, or on-prem. Documents are uploaded, read, indexed, and searched without ever leaving your network.

It replaces the OCR vendor, the workflow tool, the chat layer, and the cloud-sync middleware you pay four separate bills for today. One license covers all of it.

Your team sees your brand. So do your clients, if you resell it. Syndoc stays under the hood.

12
Document categories auto-classified
3
Cloud storage providers out of the box
1×
One-time license — no recurring fees
∞
Documents you can process

What it does

Six things your current stack can't do together.

It reasons, not just retrieves.

Agentic RAG plans a complex question into up to four sub-queries, runs them, and synthesises one grounded answer. Hybrid retrieval blends pgvector similarity with BM25 keyword ranking, so nothing relevant slips through.

It reads what humans write.

OCR on PDFs, scanned paper, Word files, and images via Azure Document Intelligence. It pulls out structured fields — invoice totals, dates, party names — and puts a confidence score on every one so your team reviews exceptions, not everything.

It files itself.

Every document is sorted into one of twelve canonical categories the moment it lands — no LLM call, no manual tagging. Filter an entire library by type in a click, and let classification drive what happens next.

It tells two versions apart.

Pick any two processed documents and Syndoc streams a structured comparison — what changed, what is new, what is missing. Point it at a specific clause or field type when you only care about one thing.

Documents that move themselves.

Build approval chains with ordered steps and assigned reviewers. Field-based triggers fire on their own — when an extracted amount clears a threshold or a document is classified as an NDA — and signed webhooks notify whatever comes next.

Drop it onto any site.

A single script tag adds a branded chat widget — your colours, your logo — to any page, backed by your own document library. Bring your cloud storage too: Google Drive, Dropbox, and OneDrive sync with your own OAuth keys.

The pipeline

A document goes in. Intelligence comes out.

The document pipeline

  1. 1

    Upload or sync

    Drag-and-drop, POST to the API, or let folders sync automatically from Google Drive, Dropbox, or OneDrive.

  2. 2

    Extract & classify

    A background worker runs OCR, pulls fields and tables with confidence scores, and sorts the document into one of 12 categories.

  3. 3

    Index

    Chunks are embedded into your vector database (pgvector or Qdrant) and keyword-indexed, ready for hybrid search.

The intelligence layer

  1. 4

    Ask

    A question arrives from the chat UI, the API, or the embeddable widget on your own site.

  2. 5

    Plan & retrieve

    Agentic mode decomposes a complex question into sub-queries, runs hybrid BM25 + vector retrieval, and can call structured document tools.

  3. 6

    Synthesise & stream

    A grounded, source-cited answer streams back token by token — with the reasoning steps shown live if you choose to expose them.

All of this runs in your environment. The worker process is separate from the API — your uploads never queue through our servers.

For builders

Ship it as your product.

Run document intelligence for your own business units, or resell it to clients as your product. Each tenant gets its own logo, colours, font, and custom domain.

Feature flags turn capabilities on per client, and the embeddable chat widget carries their brand colours. They never see the word Syndoc — unless you decide to show them.

Meridian Legal
  • Documents
  • Workflows
  • Search
  • Settings

Ownership

Buy it once. Run it forever.

Runs on your infra

Azure, AWS, or bare metal. PostgreSQL with pgvector. The storage and queue providers are configurable. You provision it, you control it.

One PO. No renewal.

Not a subscription. When you buy Syndoc, you own it. No per-seat pricing, no usage caps, no surprise invoice in month 13.

Your documents stay yours

Syndoc doesn't phone home. No telemetry, no document scanning on our end. Your AI calls go straight from your server to your own LLM provider.

Security

Built for teams where a breach isn't an option.

  • Tenant data isolated at the query level
  • OAuth credentials encrypted at rest (Fernet)
  • httpOnly session cookies, server-side validation
  • Full document audit trail — who, what, when, where
  • Soft deletes — nothing is ever hard-deleted
  • Rate limiting, X-Request-ID tracing, structured logging

We wrote this for industries where compliance isn't optional.

Who it's for

Built for teams that live inside documents.

From claims adjusters to legal associates — if your team processes high volumes of documents, Syndoc was built for you.

See all use cases →

Works with

Connects to the tools you already use.

Cloud storage — live today, bring your own keys

Google Drive Dropbox OneDrive Azure Blob AWS S3

AI backbone — configurable per deployment

Azure Document Intelligence Azure OpenAI AWS Bedrock pgvector Qdrant

Coming soon

DocuSign soonSlack soonZapier soon

Pricing

One price. No surprises.

One-Time

Syndoc License

Talk to us

Priced for your team size and deployment.

  • Agentic RAG chat — query planner, hybrid search, cited answers
  • OCR + structured field extraction with per-field confidence scores
  • Auto-classification, document comparison & field-triggered workflows
  • Embeddable chat widget + cloud storage sync (Drive, Dropbox, OneDrive)
  • Unlimited documents and users — self-hosted on your infrastructure
  • White-label rights, signed webhooks & full audit trail
  • One year of updates + priority support
  • Migration assistance
Book a Demo

Optional Add-On

Managed Deployment

Ask about it

We stand it up on your cloud account.

  • We set it up on your Azure / AWS account
  • CI/CD pipeline included
  • Monitoring + alerting configured
  • Ongoing deployment support
Get in touch

We don't do free trials — we do real demos with your actual use case.

Questions

The things people ask before they book.

Ready to stop renting your document stack?

We'll show you the platform running on a real deployment — not a recorded video.

syndoc.dev · Built for teams who need to own their tools.

Privacy Policy