[Integrations](https://spiral.ai/integrations) › [opencode](https://spiral.ai/integrations/opencode) › [SAM 3.1 Image to RLE](https://spiral.ai/integrations/opencode/sam-3-1-image-to-rle)

# How to Use SAM 3.1 Image to RLE in opencode

SAM 3.1 Image to RLE finds objects in an existing image and returns their segmentation masks as RLE data. It fits annotation pipelines and editing tools that need object selections they can process.

|  |  |
| --- | --- |
| Model | SAM 3.1 Image to RLE |
| Available through | fal |
| Category | Image |
| Price | $0.013 per run at the default inputs. |
| Returns | image |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with SAM 3.1 Image to RLE in opencode

- Use SAM 3.1 Image to RLE on the product photos in this project's sample assets with the description 'shoe.' Decode the masks and add a review page that overlays each selection on its source image.
- Add an annotation review flow using SAM 3.1 Image to RLE for 'bicycle' in the project's sample images. Keep each object selection separate and require approval before saving an annotation.
- Use SAM 3.1 Image to RLE to select 'chair' in our editor's sample photo. Connect the mask to a preview that changes only the selected region's color.
- Build a point correction control around SAM 3.1 Image to RLE. Keep correction points tied to the selected object and display the updated mask over the source image.
- Use SAM 3.1 Image to RLE to prepare face selections from this project's demo photos. Add a separate pixelation step and a review screen showing the full image before export.

## How to set up SAM 3.1 Image to RLE in opencode

Pick the way that suits you. Each takes about two minutes.

### Use Spiral's MCP server

1. **Add the server**: Leave out `--global` to add Spiral to the current project only. No header and no client ID or secret: opencode registers itself with Spiral.

```bash
opencode mcp add spiral --global --url https://api.spiral.ai/mcp
```

2. **Sign in**: Your browser opens. Sign in to Spiral, pick the organization and click Authorize, then return to the terminal. No API key. Inside opencode, `/mcps` starts the same sign-in.

```bash
opencode mcp auth spiral
```

3. **Check it**: Spiral shows as connected.

```bash
opencode mcp list
```

### Use the Spiral skill

1. **Send opencode this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run SAM 3.1 Image to RLE**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "fal-ai/sam-3-1/image-rle",
    "input": {
      "image_url": "https://raw.githubusercontent.com/facebookresearch/segment-anything-2/main/notebooks/images/truck.jpg"
    }
  }'
```

Then paste one of the prompts above: opencode finds SAM 3.1 Image to RLE on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/opencode

## SAM 3.1 Image to RLE tools and prices

- [Sam 3 1 · Image to RLE](https://spiral.ai/tools/fal/fal-ai/sam-3-1/image-rle) (fal): $0.013 per run at the default inputs.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| prompt |  | wheel | Text prompt for segmentation |
| point_prompts |  | [] | List of point prompts |
| box_prompts |  | [] | Box prompt coordinates (x_min, y_min, x_max, y_max). Multiple boxes supported - use object_id to group boxes for the same object or leave empty for separate objects. |
| apply_mask |  | true | Apply the mask on the image. |
| output_format | jpeg, png, webp | png | The format of the generated image. |
| return_multiple_masks |  | false | If True, upload and return multiple generated masks as defined by `max_masks`. |

## What can you do with SAM 3.1 Image to RLE in opencode?

Build object selection and annotation features around SAM 3.1 masks. A coding agent can connect the returned RLE data to your project's preview, correction, and export code.

### Prepare product cutouts

Use SAM 3.1 Image to RLE to select a product in a photograph, then pass the decoded mask to an editor for background removal. Review the selection before exporting the cutout.

### Draft dataset annotations

Generate object masks from short category labels with SAM 3.1 Image to RLE. Present each selection for review before adding it to a training dataset.

### Prepare selective photo edits

Select an object with SAM 3.1 Image to RLE, then use its mask in a separate editing step to recolor, blur, or highlight that region.

### Prepare privacy redactions

Find faces, license plates, or screens with SAM 3.1 Image to RLE. Apply pixelation in an editor and inspect the full image for missed regions before sharing.

### Separate objects for a collage

Request individual object selections from SAM 3.1 Image to RLE, then turn reviewed masks into cutouts for a collage or presentation.

### Build an interactive selection tool

Let a person describe an object, inspect the mask, and refine the selection with points. Use SAM 3.1 Image to RLE for segmentation and application code for previews and editing.

## Overview

- [fal](https://spiral.ai/providers/fal)
- [All opencode guides](https://spiral.ai/integrations/opencode)

## SAM 3.1 Image to RLE in other agents

- [Amp](https://spiral.ai/integrations/amp/sam-3-1-image-to-rle)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/sam-3-1-image-to-rle)
- [Claude](https://spiral.ai/integrations/claude-desktop/sam-3-1-image-to-rle)
- [Claude Code](https://spiral.ai/integrations/claude-code/sam-3-1-image-to-rle)
- [Cline](https://spiral.ai/integrations/cline/sam-3-1-image-to-rle)
- [Codex](https://spiral.ai/integrations/codex/sam-3-1-image-to-rle)
- [Cursor](https://spiral.ai/integrations/cursor/sam-3-1-image-to-rle)
- [Devin](https://spiral.ai/integrations/devin/sam-3-1-image-to-rle)
- [dots](https://spiral.ai/integrations/dots/sam-3-1-image-to-rle)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/sam-3-1-image-to-rle)
- [Gemini](https://spiral.ai/integrations/gemini/sam-3-1-image-to-rle)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/sam-3-1-image-to-rle)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/sam-3-1-image-to-rle)
- [goose](https://spiral.ai/integrations/goose/sam-3-1-image-to-rle)
- [Grok](https://spiral.ai/integrations/grok/sam-3-1-image-to-rle)
- [Grok Build](https://spiral.ai/integrations/grok-build/sam-3-1-image-to-rle)
- [Hermes](https://spiral.ai/integrations/hermes/sam-3-1-image-to-rle)
- [Instinct](https://spiral.ai/integrations/instinct/sam-3-1-image-to-rle)
- [Manus](https://spiral.ai/integrations/manus/sam-3-1-image-to-rle)
- [Muse](https://spiral.ai/integrations/muse/sam-3-1-image-to-rle)
- [Muse Code](https://spiral.ai/integrations/muse-code/sam-3-1-image-to-rle)
- [n8n](https://spiral.ai/integrations/n8n/sam-3-1-image-to-rle)
- [Notion](https://spiral.ai/integrations/notion/sam-3-1-image-to-rle)
- [OpenClaw](https://spiral.ai/integrations/openclaw/sam-3-1-image-to-rle)
- [Perplexity](https://spiral.ai/integrations/perplexity/sam-3-1-image-to-rle)
- [Raycast](https://spiral.ai/integrations/raycast/sam-3-1-image-to-rle)
- [VS Code](https://spiral.ai/integrations/vscode/sam-3-1-image-to-rle)
- [Windsurf](https://spiral.ai/integrations/windsurf/sam-3-1-image-to-rle)
- [Zapier](https://spiral.ai/integrations/zapier/sam-3-1-image-to-rle)

## More opencode guides

- [SAM 3.1 Image](https://spiral.ai/integrations/opencode/sam-3-1-image)
- [Sana](https://spiral.ai/integrations/opencode/sana)
- [SAM 3 Image](https://spiral.ai/integrations/opencode/sam-3-image)
- [Sana 1.5 1.6B](https://spiral.ai/integrations/opencode/sana-1-5-1-6b)
- [RunDiffusion Photo Flux](https://spiral.ai/integrations/opencode/rundiffusion-photo-flux)
- [Sana 1.5 4.8B](https://spiral.ai/integrations/opencode/sana-1-5-4-8b)

## Questions about SAM 3.1 Image to RLE in opencode

**Can opencode use SAM 3.1 Image to RLE?**

Yes. Add Spiral to opencode as an MCP server, and opencode can find and run SAM 3.1 Image to RLE like any built-in tool.

**How much does SAM 3.1 Image to RLE cost in opencode?**

$0.013 per run at the default inputs. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your opencode plan is separate.

**Do I need a fal account to use SAM 3.1 Image to RLE in opencode?**

No. Spiral runs SAM 3.1 Image to RLE on its own fal account. You sign in to Spiral once and never paste a key.

**How long does it take to set up SAM 3.1 Image to RLE in opencode?**

About two minutes: add the server, then sign in, then check it.

**Does opencode wait while SAM 3.1 Image to RLE runs?**

SAM 3.1 Image to RLE runs in the background. opencode starts the run, checks on it and gives you the image when it's ready.

**What does SAM 3.1 Image to RLE produce?**

It returns object segmentation masks encoded with run-length encoding, or RLE. An agent can use that mask data in an annotation tool or decode it to prepare an editing selection.

**How should I describe the object I want?**

Start with a short noun phrase such as "wheel," "red umbrella," or "hardcover book." Keep the request focused on an object and one or two useful attributes.

**Can I ask for an object by its position or relationship to other objects?**

Long referring expressions exceed the documented direct prompting scope. Have the agent identify the intended object first, then use a short description or a visual selection to guide segmentation.

**Can SAM 3.1 select several objects of the same kind?**

Concept segmentation is designed to find all instances matching a description. Request separate masks when your next step needs individual objects, then review the selections for missed or unwanted matches.

**How can I correct a selection?**

Use point guidance to refine the intended object. Keep points associated with the same object identifier when they should update one selection.

