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

# Changelog

> Every user-visible change to the Overten API.

The full engineering changelog lives in the backend repo
(`backend/docs/api/CHANGELOG.md`). This page mirrors the user-facing
highlights.

We follow [Keep a Changelog](https://keepachangelog.com) and
semantic-ish versioning — breaking changes always ship under a new URL
prefix (`/api/v2/*`), and `/api/v1/*` stays live for at least 12 months
after any major version bump.

***

## Unreleased (v1.0.0-alpha)

The initial API surface. Everything below is new.

### Endpoints

* **Self-service signup** — `POST /signup` creates a workspace org and
  your first API key in one call (Firebase-authed).
* **Organization key management** —
  `POST / GET / DELETE /organizations/{org_id}/api-keys`.
* **Identity** — `GET /verify` validates a key and returns org context.
* **Preflight** — `POST /preflight` dry-runs a request through the LLM
  validator without spending credits.
* **File uploads** — `POST /files` for generic files.
* **Image uploads** — `POST /images` with tags/caption, plus full CRUD
  (`GET /images`, `GET /images/{id}`, `DELETE /images/{id}`).
* **Generation (natural language)** — `POST /excel/generate`,
  `POST /word/generate`, `POST /slides/generate`. Sync (default) or
  `async: true`. Resume via `run_id`.
* **Excel analysis** — `POST /excel/analyze` returns a JSON answer
  against an uploaded workbook.
* **Slide outlines** — `POST /slides/outline` for preview-before-generate
  workflows.
* **Tasks** — `GET /tasks/{task_id}` for async polling.
* **Runs** — `GET /runs`, `GET /runs/{id}`, `GET /runs/{id}/download`,
  `POST /runs/{id}/export` (PDF / CSV / PNG / etc.).
* **Webhooks** — `POST / GET / DELETE /webhooks` for subscriptions,
  plus per-request `webhook_url` on generation endpoints.
* **Embedded editor** — every completed run includes an `edit_url` your
  app can iframe.

### Features

* **Sync and async modes** on every generation endpoint — sync promotes
  to async automatically at the 300s ceiling.
* **Run resume semantics** — pass `run_id` to iterate on the same
  document without re-explaining what it is.
* **Preflight LLM validator** — rejects requests referencing missing
  assets before spending real credits.
* **HMAC-signed webhooks** — SHA-256, replay-protection timestamp, 7
  retry attempts with exponential backoff.
* **Credit ledger** — every movement logged for audit.
* **Org-scoped tenancy** — every resource lookup enforces
  `(caller_org == resource_org)` — no cross-tenant leaks.

### Supported formats

* Excel: `.xlsx` primary, exports to pdf / csv / png / html
* Word: `.docx` primary, exports to pdf / html / txt / odt
* Slides: `.pptx` primary, exports to pdf / png / odp

### Current limitations

* **Personal vs. workspace key scoping** — available via the dashboard
  key-creation flow; every key inherits the creator's current workspace
  role.
* **Text and static visuals only** — no audio or video generation.
* **Slides: max 30 slides per call.** Larger decks should be built in
  chunks via `run_id` resume.
* **Webhooks fire on `run.completed` and `run.failed`** only.

***

## Versioning policy

* **URL-based**: breaking changes land at `/api/v2/*`; `/api/v1/*` stays
  live for at least 12 months.
* **Additive changes** (new endpoints, new optional fields, new enum
  values, new error codes) are always safe within v1.
* **Breaking changes** require a deprecation window:
  * Announce here + email notice.
  * Response includes `X-Deprecated: true` and `Sunset: <date>` headers
    for 12+ months.
  * Removal only on the v2 boundary.

## Follow releases

* Follow [@overten](https://x.com/overtenai) on X.
* Join our [community](https://overten.ai/community).