[Integrations](https://spiral.ai/integrations) › [opencode](https://spiral.ai/integrations/opencode) › [Moondream 2 object bounding boxes](https://spiral.ai/integrations/opencode/moondream-2-object-bounding-boxes)

# How to Use Moondream 2 object bounding boxes in opencode

Returns bounding boxes around matching objects in an image, plus an annotated image. Use it to outline products, prepare crops, or review where objects appear.

|  |  |
| --- | --- |
| Model | Moondream 2 object bounding boxes |
| Available through | fal |
| Category | Data |
| Price | $0.026 per image. |
| Returns | json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with Moondream 2 object bounding boxes in opencode

- Use Moondream 2 object bounding boxes on a public photo of bicycles outside Amsterdam Centraal, then add an overlay in this project showing the returned bicycle boxes.
- Use Moondream 2 object bounding boxes on a public photo of the Golden Gate Bridge to locate its towers, then add a crop preview for each returned box.
- Use Moondream 2 object bounding boxes on a public IKEA room photo to locate chairs, then add a review panel showing the annotated image beside the detection list.
- Use Moondream 2 object bounding boxes to look this up and give me the result as a table. Save the result in ./assets.
- Pull the latest numbers from Moondream 2 object bounding boxes and tell me what changed since last week.

## How to set up Moondream 2 object bounding boxes 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 Moondream 2 object bounding boxes**: 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/moondream2/object-detection",
    "input": {
      "image_url": "https://llava-vl.github.io/static/images/monalisa.jpg",
      "object": "<object>"
    }
  }'
```

Then paste one of the prompts above: opencode finds Moondream 2 object bounding boxes on Spiral, shows the price and runs it.

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

## Moondream 2 object bounding boxes tools and prices

- [Moondream2 · Object-Detection](https://spiral.ai/tools/fal/fal-ai/moondream2/object-detection) (fal): $0.026 per image.

## What can you do with Moondream 2 object bounding boxes in opencode?

Build an image review feature that draws Moondream locations over product photos or screenshots. A coding agent can turn boxes into outlines or points into numbered markers in your project.

### Annotate product photos

Locate shoes, bags, or furniture in catalog images. Use boxes to outline each detected product or points to place numbered review markers.

### Prepare visual counts

Find repeated objects such as bicycles, apples, or chairs and count the returned locations. Show the annotations beside the total so a reviewer can inspect the detections.

### Locate interface controls

Find a search button or cart icon in a screenshot. Use its box to highlight the control or its point location to propose a click target.

### Create draft image annotations

Generate initial boxes or point labels for a collection of object photos. Present the results for review before adding them to an annotation dataset.

### Compare scene snapshots

Locate the same object category in successive shelf or display photos. Compare the detected positions and annotated images to prepare a report of visible changes.

### Build illustrated object guides

Mark recognizable objects in museum, landmark, or equipment photos. Use boxes for a visual index or points for numbered callouts linked to explanatory text.

## Overview

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

## Moondream 2 object bounding boxes in other agents

- [Amp](https://spiral.ai/integrations/amp/moondream-2-object-bounding-boxes)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/moondream-2-object-bounding-boxes)
- [Claude](https://spiral.ai/integrations/claude-desktop/moondream-2-object-bounding-boxes)
- [Claude Code](https://spiral.ai/integrations/claude-code/moondream-2-object-bounding-boxes)
- [Cline](https://spiral.ai/integrations/cline/moondream-2-object-bounding-boxes)
- [Codex](https://spiral.ai/integrations/codex/moondream-2-object-bounding-boxes)
- [Cursor](https://spiral.ai/integrations/cursor/moondream-2-object-bounding-boxes)
- [Devin](https://spiral.ai/integrations/devin/moondream-2-object-bounding-boxes)
- [dots](https://spiral.ai/integrations/dots/moondream-2-object-bounding-boxes)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/moondream-2-object-bounding-boxes)
- [Gemini](https://spiral.ai/integrations/gemini/moondream-2-object-bounding-boxes)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/moondream-2-object-bounding-boxes)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/moondream-2-object-bounding-boxes)
- [goose](https://spiral.ai/integrations/goose/moondream-2-object-bounding-boxes)
- [Grok](https://spiral.ai/integrations/grok/moondream-2-object-bounding-boxes)
- [Grok Build](https://spiral.ai/integrations/grok-build/moondream-2-object-bounding-boxes)
- [Hermes](https://spiral.ai/integrations/hermes/moondream-2-object-bounding-boxes)
- [Instinct](https://spiral.ai/integrations/instinct/moondream-2-object-bounding-boxes)
- [Manus](https://spiral.ai/integrations/manus/moondream-2-object-bounding-boxes)
- [Muse](https://spiral.ai/integrations/muse/moondream-2-object-bounding-boxes)
- [Muse Code](https://spiral.ai/integrations/muse-code/moondream-2-object-bounding-boxes)
- [n8n](https://spiral.ai/integrations/n8n/moondream-2-object-bounding-boxes)
- [Notion](https://spiral.ai/integrations/notion/moondream-2-object-bounding-boxes)
- [OpenClaw](https://spiral.ai/integrations/openclaw/moondream-2-object-bounding-boxes)
- [Perplexity](https://spiral.ai/integrations/perplexity/moondream-2-object-bounding-boxes)
- [Raycast](https://spiral.ai/integrations/raycast/moondream-2-object-bounding-boxes)
- [VS Code](https://spiral.ai/integrations/vscode/moondream-2-object-bounding-boxes)
- [Windsurf](https://spiral.ai/integrations/windsurf/moondream-2-object-bounding-boxes)
- [Zapier](https://spiral.ai/integrations/zapier/moondream-2-object-bounding-boxes)

## More opencode guides

- [Mistral document text extraction](https://spiral.ai/integrations/opencode/mistral-document-text-extraction)
- [Moondream 2 object coordinates](https://spiral.ai/integrations/opencode/moondream-2-object-coordinates)
- [Mistral document text and table extraction](https://spiral.ai/integrations/opencode/mistral-document-text-and-table-extraction)
- [SAM 3 image embedding](https://spiral.ai/integrations/opencode/sam-3-image-embedding)
- [Mistral document OCR](https://spiral.ai/integrations/opencode/mistral-document-ocr)
- [Sa2VA 4B image segmentation](https://spiral.ai/integrations/opencode/sa2va-4b-image-segmentation)

## Questions about Moondream 2 object bounding boxes in opencode

**Can opencode use Moondream 2 object bounding boxes?**

Yes. Add Spiral to opencode as an MCP server, and opencode can find and run Moondream 2 object bounding boxes like any built-in tool.

**How much does Moondream 2 object bounding boxes cost in opencode?**

$0.026 per image. 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 Moondream 2 object bounding boxes in opencode?**

No. Spiral runs Moondream 2 object bounding boxes on its own fal account. You sign in to Spiral once and never paste a key.

**How long does it take to set up Moondream 2 object bounding boxes in opencode?**

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

**Does opencode wait while Moondream 2 object bounding boxes runs?**

Moondream 2 object bounding boxes runs in the background. opencode starts the run, checks on it and gives you the json when it's ready.

**What moment does a Moondream result describe?**

It describes objects in the supplied still image. Moondream documents video detection as a frame-by-frame workflow, so use a new frame to examine a later moment.

**What is the difference between boxes and points?**

Moondream object detection returns rectangular boundaries around objects. Pointing returns center coordinates for matching instances, which suits markers and position-based actions.

**Are point coordinates measured in pixels?**

Moondream Point returns x,y coordinates normalized between 0 and 1. Multiply x by the image width and y by its height when you need pixel positions.

**Can Moondream locate more than one matching object?**

Yes. Detection supports multiple matching objects in one query, and pointing returns a center location for each matching instance.

**Am I limited to standard object categories?**

Moondream supports descriptions beyond a predefined category list. Pointing also accepts features and spatial relationships, so a request such as 'the handle on the left cup' can be more useful than 'cup.'

