Skip to main content
The Overten API doesn’t just edit files. It understands your entire project. Upload a set of documents and the API cross-references them, extracts facts, catches conflicts before they become problems, and lets you surface the result as structured insights or a polished report. And when you need to produce a document from scratch — a spreadsheet, a report, a deck — the same API generates .xlsx, .docx, and .pptx files on demand, from a single prompt.

Quickstart

Create a project, attach files, read insights — in five minutes.

Get an API key

Sign up and mint a sk_live_* key in one request.

API reference

Every endpoint, request schema, and response shape.

Changelog

Release notes and deprecations.

​
What you can build

  • Project intelligence — the primary surface. Upload a folder of files into a project and the pipeline extracts facts, reconciles them across documents, flags discrepancies, and surfaces insights. Poll for status, read the alert feed, export a branded .docx report. See the Project Intelligence guide.
  • Cross-document conflict detection — budget mismatches, deadline drift, scope discrepancies, duplicated claims. The pipeline highlights them with citations and recommended actions.
  • Multi-file reasoning — ask natural-language questions against an Excel, Word, or PowerPoint file and get structured answers back. Extract totals from a workbook, pull clauses from a contract, summarize a deck without producing a new file.
  • Document generation — describe what you want, get a polished file back. 60+ document actions across Word, Excel, and PowerPoint. Useful standalone — or paired with the intelligence layer to auto-generate briefs, reports, and decks from a project’s findings.
  • Document conversion — turn any .docx/.xlsx/.pptx you upload (or one we generated) into .pdf, .html, .txt, .csv, .png, etc. See the conversions guide.

​
How it works

  1. Sign up via a Firebase token and get an sk_live_* key.
  2. Create a project via POST /projects and attach files with POST /projects/{id}/files. The pipeline kicks off automatically.
  3. Poll status via GET /projects/{id}/analysis/status or read the insights feed via GET /projects/{id}/insights.
  4. Export a report via POST /projects/{id}/reports — a deterministic .docx grouped by severity.
  5. Need a standalone document? Generate via POST /word/generate, /excel/generate, or /slides/generate with a prompt — same signed-URL response, no project required.
Every generation and report call returns a run_id. Pass it back on the next call to resume — the agent keeps its memory of what it built, so follow-up edits cost a fraction of what turn 1 did.

​
How you call it

Create a project, attach files, read what the pipeline surfaces: 
POST /projects
{ "name": "Acme Q3 Due Diligence" }
→ { "project_id": "proj_..." }

POST /projects/{project_id}/files  (multipart with file + role)

GET /projects/{project_id}/insights?severity=critical,high
Or generate a document from a prompt: 
POST /excel/generate
{ "prompt": "Create a Q3 sales report with summary statistics" }
The same generation shape works for /word/generate and /slides/generate. See the Project Intelligence guide and format guides for what each surface can do.

​
Sync and async

Every generation + report endpoint accepts async: true. Sync returns the finished document in the HTTP response (up to a 300-second (5 min) ceiling). Async returns a task_id immediately; poll GET /tasks/{task_id} or register a webhook. See Async vs sync for when to use each.

​
Status & support