How to Use EVF-SAM2 in opencode
EVF-SAM2 turns an English description into a selection mask for an existing image. Use it to isolate a named object or select hair and background regions with semantic segmentation for later editing.
Prompts to try with EVF-SAM2 in opencode
Copy an example into opencode, or ask in your own words.
Use EVF-SAM2 on assets/product.jpg to select 'the shoe on the left.' Save the binary mask beside the original and create an overlay for review.
Use EVF-SAM2 on assets/portrait.jpg with semantic segmentation to select 'hair.' Save the mask and add a preview so I can inspect the selection.
Use EVF-SAM2 on assets/room.jpg to select 'the chair beside the window.' Prepare the mask for a later recoloring step.
Use EVF-SAM2 on assets/table.jpg to select 'the plate' and exclude 'the fork.' Show the resulting selection over the original.
Use EVF-SAM2 on the project's sample landscape image with semantic segmentation to select 'sea.' Add the mask to our image-editing preview.
How to set up EVF-SAM2 in opencode
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- 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.
opencode mcp add spiral --global --url https://api.spiral.ai/mcp
- 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.
opencode mcp auth spiral
- Check it
Spiral shows as connected.
opencode mcp list
Use the Spiral skill
- Send opencode this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- Install and sign in
Installs the Spiral CLI, signs you in and adds its skill to your coding agents.
npx -y spiral-cli@latest init
Call the Spiral API
- Create an API key
In the console, under API keys.
- Run EVF-SAM2
One request; the result comes back by polling or a webhook.
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/evf-sam", "input": { "prompt": "Cat in the middle of the image", "image_url": "https://storage.googleapis.com/falserverless/web-examples/evf-sam2/evfsam2-cat.png" } }'
Then paste one of the prompts above: opencode finds EVF-SAM2 on Spiral, shows the price and runs it.
EVF-SAM2 providers and prices
Spiral runs EVF-SAM2 through fal. Open a card for today's price, with every fee included.
Settings you can ask for
EVF-SAM2 works from your prompt and image URL. Ask opencode for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Negative prompt | what to leave out, in words | Areas to exclude from segmentation (will be subtracted from prompt results) | |
| Semantic type | on or off | off | Enable semantic level segmentation for body parts, background or multi objects |
| Mask only | on or off | on | Output only the binary mask instead of masked image |
| Use grounding dino | on or off | off | Use GroundingDINO instead of SAM for segmentation |
| Revert mask | on or off | off | Invert the mask (background becomes foreground and vice versa) |
| Blur mask | 0 to 50 | 0 | Apply Gaussian blur to the mask. Value determines kernel size (must be odd number) |
| Expand mask | 0 to 20 | 0 | Expand/dilate the mask by specified pixels |
| Fill holes | on or off | off | Fill holes in the mask using morphological operations |
Every API field of EVF-SAM2 Segmentation is on its tool page: open its card above.
What can you do with EVF-SAM2 in opencode?
Turn image assets in your project into selections for editing and annotation. Use EVF-SAM2 to describe the target in English, then build a review step around the returned mask.
Product selections for recoloring
Use EVF-SAM2 to mask a specific shoe, bag, or chair in a product photo, then pass the selection to an editing tool.
Background replacement preparation
Use EVF-SAM2 with semantic segmentation to select a photo's background before replacing it in a separate editing step.
Hair color previews
Use EVF-SAM2 with semantic segmentation to select hair for a color preview. Inspect the selection before applying the edit.
One object in a crowded photo
Use EVF-SAM2 to select a target by its location or relationship to another object, such as the cup beside a plate.
Draft annotation masks
Use EVF-SAM2 to propose object masks from English descriptions, then have an annotator review them before adding them to a dataset.
Landscape editing selections
Use EVF-SAM2 with semantic segmentation to select a region such as the sea for a later color adjustment.
More on EVF-SAM2 and opencode
Overview
EVF-SAM2 in other agents
Questions about EVF-SAM2 in opencode
Can opencode use EVF-SAM2?
Yes. Add Spiral to opencode as an MCP server, and opencode can find and run EVF-SAM2 like any built-in tool.
How much does EVF-SAM2 cost in opencode?
$0.0065 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 EVF-SAM2 in opencode?
No. Spiral runs EVF-SAM2 on its own fal account. You sign in to Spiral once and never paste a key.
How long does it take to set up EVF-SAM2 in opencode?
About two minutes: add the server, then sign in, then check it.
Does opencode wait while EVF-SAM2 runs?
EVF-SAM2 runs in the background. opencode starts the run, checks on it and gives you the image when it's ready.
What is EVF-SAM2 useful for?
Use EVF-SAM2 to select a region in an existing image by describing it. The resulting mask can serve as the selection for a later edit, such as changing a product's color or replacing a background.
How do I select one object when several look alike?
Name the object and add a visible detail or location, such as 'the red cup beside the plate.' EVF-SAM supports spatial descriptions, so a relationship to another object can help distinguish your target.
Which language should I use for EVF-SAM2 prompts?
Use English. The official demo specifies English-only support, so translate other languages before submitting the target description.
Can EVF-SAM2 select hair or a background region?
Yes. Use semantic segmentation for body parts, background regions, or multiple objects, then describe the region you want.
How do I exclude something from a selection?
Describe the region to keep in the positive description and the unwanted region in the negative description. EVF-SAM2 subtracts the negative selection from the positive result.
Start using EVF-SAM2 in opencode
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into opencode:
set up https://spiral.ai/SKILL.md


