Fotographer AI, Inc

AI-powered image generation and inpainting API supporting three generation modes:
- **Mode 1: Inpainting** - Traditional inpainting with subject, mask, and background
- **Mode 2: Scene Generation** - Generate scenes from depth maps or 3D previews
- **Mode 3: Background Generation** - Generate new backgrounds around subjects using edge detection

The API supports both **synchronous** and **asynchronous (polling)** processing modes.

---

## 📍 Important: Endpoint Structure

This OpenAPI spec uses **virtual paths** for clarity. The actual endpoints are:

**Sync Mode (all gen_modes):**
- Actual URL: `https://zen-inpaint-1066271267292.europe-west1.run.app`
- Virtual paths in this spec: `/inpaint-sync`, `/scene-generation-sync`, `/bg-generation-sync`

**Async Mode (all gen_modes):**
- Actual URL: `https://zeninpaint-pro-1066271267292.europe-west1.run.app/run`
- Virtual paths in this spec: `/inpaint-async`, `/scene-generation-async`, `/bg-generation-async`

**Polling:**
- Actual URL: `https://zeninpaint-pro-1066271267292.europe-west1.run.app/predict`

The virtual paths help organize the documentation by generation mode, but all requests to the sync server use the base URL, and all async job submissions use `/run`.

---

## 🔑 Authentication

All endpoints (except `/ping`) require two headers:
- `x-api-key`: Your API key
- `x-email`: Your registered email

## 💳 Credits

Each request consumes 0.04 credits. Ensure sufficient credits before making requests.


Zeninpaint API

Traditional inpainting combining subject, mask, and background.
Available in both sync and async modes.


Mode 1 - Inpainting

**Actual Endpoint:** `POST https://zen-inpaint-1066271267292.europe-west1.run.app`

Traditional inpainting mode that combines a subject, mask, and background image.
Returns the generated image directly in the response.

**Concurrent Requests:** Up to 14 simultaneous requests
**Timeout:** 300 seconds


Mode 1 - Inpainting (Sync)

**Actual Endpoint:** `POST https://zeninpaint-pro-1066271267292.europe-west1.run.app/run`

Submit an inpainting job asynchronously. Returns a ticket ID for polling.
Use the `/predict` endpoint to check status and retrieve results.

**Advantages:**
- No timeout issues for long-running jobs
- Process multiple requests concurrently
- Queue system manages resources automatically


Mode 1 - Inpainting (Async)

Generate photorealistic scenes from depth maps or 3D previews.
Available in both sync and async modes.


Mode 2 - Scene Generation

**Actual Endpoint:** `POST https://zen-inpaint-1066271267292.europe-west1.run.app`

Generate photorealistic scenes from depth maps or 3D preview renders.
The system extracts depth maps and canny structure from preview images.

**Concurrent Requests:** Up to 14 simultaneous requests
**Timeout:** 300 seconds


Mode 2 - Scene Generation (Sync)

**Actual Endpoint:** `POST https://zeninpaint-pro-1066271267292.europe-west1.run.app/run`

Submit a scene generation job asynchronously. Returns a ticket ID for polling.
Use the `/predict` endpoint to check status and retrieve results.


Mode 2 - Scene Generation (Async)

Generate new backgrounds around subjects using edge detection.
Requires PNG with transparent background.
Available in both sync and async modes.


Mode 3 - Background Generation

**Actual Endpoint:** `POST https://zen-inpaint-1066271267292.europe-west1.run.app`

Generate new realistic backgrounds around a subject while maintaining its placement.
Uses edge detection to preserve subject boundaries.

**Subject Requirement:** PNG with transparent background
**Concurrent Requests:** Up to 14 simultaneous requests
**Timeout:** 300 seconds


Mode 3 - Background Generation (Sync)

**Actual Endpoint:** `POST https://zeninpaint-pro-1066271267292.europe-west1.run.app/run`

Submit a background generation job asynchronously. Returns a ticket ID for polling.
Use the `/predict` endpoint to check status and retrieve results.


Mode 3 - Background Generation (Async)

Check status and retrieve results for async jobs

Polling

**Actual Endpoint:** `POST https://zeninpaint-pro-1066271267292.europe-west1.run.app/predict`

Check the status of an asynchronous job and retrieve results when ready.
Works with all generation modes (inpaint, scene_generation, bg_generation).

**Polling Best Practices:**
- Poll every 1-2 seconds for optimal performance
- Set max wait time to 300 seconds (5 minutes)
- Handle network errors with retry logic


Poll Job Status

Health

Ping endpoint to wake up cold instances and set priority for faster subsequent requests.
Does NOT require authentication.


Ping & Warm-up

Alternative POST method for ping endpoint.
Does NOT require authentication.


Ping & Warm-up (POST)

Default

Zen-Inpaint

Sync Mode Server

Async Mode Server (Polling)

<p class="slate-p " >AI-powered image generation and inpainting API supporting three generation modes:</p><ul class="slate-ul " ><li class="slate-li " ><div style="position:relative"><strong class="slate-bold">Mode 1: Inpainting</strong> - Traditional inpainting with subject, mask, and background</div></li><li class="slate-li " ><div style="position:relative"><strong class="slate-bold">Mode 2: Scene Generation</strong> - Generate scenes from depth maps or 3D previews</div></li><li class="slate-li " ><div style="position:relative"><strong class="slate-bold">Mode 3: Background Generation</strong> - Generate new backgrounds around subjects using edge detection</div></li></ul><p class="slate-p " >The API supports both <strong class="slate-bold">synchronous</strong> and <strong class="slate-bold">asynchronous (polling)</strong> processing modes.</p><div class="slate-hr"><hr contenteditable="false" class="slate-HrElement-hr"/></div><h2 class="slate-h2 " >📍 Important: Endpoint Structure</h2><p class="slate-p " >This OpenAPI spec uses <strong class="slate-bold">virtual paths</strong> for clarity. The actual endpoints are:</p><p class="slate-p " ><strong class="slate-bold">Sync Mode (all gen_modes):</strong></p><ul class="slate-ul " ><li class="slate-li " ><div style="position:relative">Actual URL: <code class="slate-code">https://zen-inpaint-1066271267292.europe-west1.run.app</code></div></li></ul><p class="slate-p " ><strong class="slate-bold">Async Mode (all gen_modes):</strong></p><ul class="slate-ul " ><li class="slate-li " ><div style="position:relative">Actual URL: <code class="slate-code">https://zeninpaint-pro-1066271267292.europe-west1.run.app/run</code></div></li></ul><p class="slate-p " ><strong class="slate-bold">Polling:</strong></p><ul class="slate-ul " ><li class="slate-li " ><div style="position:relative">Actual URL: <code class="slate-code">https://zeninpaint-pro-1066271267292.europe-west1.run.app/predict</code></div></li></ul><p class="slate-p " >The virtual paths help organize the documentation by generation mode, but all requests to the sync server use the base URL, and all async job submissions use <code class="slate-code">/run</code>.</p><div class="slate-hr"><hr contenteditable="false" class="slate-HrElement-hr"/></div><h2 class="slate-h2 " >🔑 Authentication</h2><p class="slate-p " >All endpoints (except <code class="slate-code">/ping</code>) require two headers:</p><ul class="slate-ul " ><li class="slate-li " ><div style="position:relative"><code class="slate-code">x-api-key</code>: Your API key</div></li><li class="slate-li " ><div style="position:relative"><code class="slate-code">x-email</code>: Your registered email</div></li></ul><h2 class="slate-h2 " >💳 Credits</h2><p class="slate-p " >Each request consumes 0.04~ <a href="https://0.05" target="_self">0.05</a> (for &gt;2k images) credits. Ensure sufficient credits before making requests.</p><p class="slate-p " ></p>

AI-powered image generation and inpainting API supporting three generation modes:

The virtual paths help organize the documentation by generation mode, but all requests to the sync server use the base URL, and all async job submissions use 

 (for >2k images) credits. Ensure sufficient credits before making requests.

<p class="slate-p " >Traditional inpainting combining subject, mask, and background.
Available in both sync and async modes.</p><p class="slate-p " ></p>

Traditional inpainting combining subject, mask, and background.
Available in both sync and async modes.

<p class="slate-p " >Generate new backgrounds around subjects using edge detection.
Requires PNG with transparent background.
Available in both sync and async modes.</p><p class="slate-p " ></p>

Generate new backgrounds around subjects using edge detection.
Requires PNG with transparent background.
Available in both sync and async modes.

<p class="slate-p " >Check status and retrieve results for async jobs</p><p class="slate-p " ></p>

<p class="slate-p " >Health check and warm-up endpoints</p><p class="slate-p " ></p>

<p class="slate-p " >Ping endpoint to wake up cold instances and set priority for faster subsequent requests.
Does NOT require authentication.</p><p class="slate-p " ></p>

Ping endpoint to wake up cold instances and set priority for faster subsequent requests.
Does NOT require authentication.

status

message

<p class="slate-p " >Alternative POST method for ping endpoint.
Does NOT require authentication.</p><p class="slate-p " ></p>

Alternative POST method for ping endpoint.
Does NOT require authentication.

Sections