OpenClaw-only documentation

The OpenClaw skill page for using Pop-cam the right way

This page is written only for OpenClaw. It covers every feature the Pop-cam skill offers — image-to-image, text-to-image, async webhooks, custom prompts, and credits — plus the full onboarding guide for sign-up, token creation, and authenticated API usage.

This page is only for OpenClaw users.

OpenClaw users must create a Pop-cam account before using the skill.

After sign-in, create a Pop-cam API token and use it in OpenClaw.

Supports both image-to-image and text-to-image generation via a single endpoint.

Open sign-up Open sign-in Open developer settings

OpenClaw Flow

Watch the handoff step by step

Auto-playing walkthrough

Step

OpenClaw cannot use the Pop-cam skill until the user has a Pop-cam account. The first instruction should always be to create an account if the user does not already have one.

Open https://www.pop-cam.com/sign-up.

Create the account with the user’s preferred email sign-up flow.

Finish any required verification step before moving on.

Step

If the user is not signed in, OpenClaw should stop and prompt them to sign in. Pop-cam token creation and account-linked usage depend on an authenticated session.

Open https://www.pop-cam.com/sign-in.

Wait until the user is signed in successfully.

Do not continue to token creation until sign-in is complete.

Step

Generate Your API Token

Once signed in, OpenClaw should direct the user to the Developer Settings page to create a Pop-cam API token. The token is shown once, so the user should save it immediately.

Open https://www.pop-cam.com/developer.

Create a new token with a clear label such as OpenClaw.

Copy the token immediately because the full token is not shown again.

Step

Use The Token Inside OpenClaw

The Pop-cam skill should send the user’s token as a Bearer token on authenticated requests. Each authenticated generation uses the user’s Pop-cam credits.

Send Authorization: Bearer pk_... on Pop-cam API requests.

Use the token only for the signed-in user who created it.

If the token fails, ask the user to revoke it and generate a new one.

Step

Generate Images From Text Prompts

OpenClaw can generate images without a source photo by sending only a prompt field and omitting the image field. This is the text-to-image mode. It costs the same 1 credit per generation as image-to-image.

Omit the image field from the request body entirely.

Provide a descriptive prompt field describing the image to create.

The response includes mode: text-to-image so OpenClaw can confirm which path was used.

ClawHub Registry

Install Pop-cam from ClawHub

The Pop-cam skill is published on ClawHub — OpenClaw's public skill registry. Install it directly or let OpenClaw discover it automatically.

🍌

pop-cam

v1.2.0 · by @petagit

AI image generation — transform photos or generate images from text prompts via the Pop-cam NanoBanana API. Supports sync and async webhook modes, custom prompts, and credit-based usage.

View on ClawHub clawhub.ai/skills/pop-cam

Quick Install

Install the skill into your local OpenClaw workspace with a single command. Updates are published to ClawHub whenever new features or fixes are added.

terminal

clawhub install pop-cam

update to latest

clawhub update pop-cam

Requires the clawhub CLI. Install with npm i -g clawhub.

Skill Features

Everything the Pop-cam skill can do

These are the capabilities OpenClaw gains when using the Pop-cam skill. Each feature is available through a single API endpoint with the user's pk_ token.

Image-to-Image Generation

1 credit

Transform any photo based on your prompt. The output can be photorealistic, cinematic, illustrated, cartoon-style, or any other look you describe — the model adapts to whatever you ask for.

Send the image field as a base64 data URL (data:image/png;base64,...).

Include a prompt describing the desired output — photorealistic enhancement, artistic transformation, cartoon conversion, etc.

Works with PNG, JPEG, and WebP inputs.

Text-to-Image Generation

1 credit

Generate an image from a text description alone — no source photo needed. Supports any visual style: photorealistic portraits, product shots, illustrations, concept art, and more.

Omit the image field entirely from the request body.

The prompt field is required — describe exactly what you want, including the visual style (e.g. "photorealistic", "cinematic", "watercolor").

The response includes mode: "text-to-image" to confirm the generation path.

Webhook / Async Mode

For workflows that cannot block on a response, provide a webhook_url. The API returns immediately with a job_id and POSTs the result to your URL when complete.

Include webhook_url in the request body with a publicly reachable HTTPS endpoint.

The API responds with HTTP 202 and a job_id instantly.

Pop-cam retries delivery up to 3 times with exponential backoff if your endpoint is unavailable.

Works with both image-to-image and text-to-image modes.

Synchronous Mode

The default mode. The API blocks until the image is generated, then returns the result directly in the response body.

Simply omit the webhook_url field.

Response includes image_url, used_prompt, credits_remaining, and mode.

Typical response time is 5–15 seconds depending on complexity.

Custom Prompt Control

Pass a prompt field to control exactly how the image is generated. Describe any visual style — photorealistic, cinematic, cartoon, watercolor, 3D render, sketch, and more.

The prompt field accepts free-form text up to the model's token limit.

For image-to-image, the prompt guides how the source image is transformed.

For text-to-image, the prompt is the sole creative input.

Supported styles include: photorealistic, cinematic, illustration, cartoon, oil painting, watercolor, 3D render, pencil sketch, anime, and any other style you can describe.

If omitted during image-to-image, a high-quality default prompt is used.

Credit-Based Usage

Pay-per-use

Every generation costs 1 credit. New accounts start with free credits. Additional credits can be purchased through Pop-cam's checkout page.

Each image generation (both modes) costs exactly 1 credit.

New Pop-cam accounts receive starter credits for free.

Every API response includes credits_remaining so OpenClaw can warn the user before they run out.

Credits can be purchased at https://www.pop-cam.com/checkout.

API Reference

Endpoints the OpenClaw skill should call

The complete list of endpoints available to OpenClaw. The generation endpoint uses a pk_ API token for authentication. The auth-status endpoint is unauthenticated and should be called first to check whether the user needs to sign in.

GET/api/openclaw/skillAuth: none

Returns the complete Pop-cam skill definition as structured JSON. OpenClaw should call this first to discover all available features, endpoints, code examples, authentication requirements, and credit costs. No auth needed.

Response

Field	Type	Description
name	string	Skill name.
version	string	Skill definition version.
description	string	Human-readable summary of what the skill does.
features	array	All available features (image-to-image, text-to-image, webhook, etc.).
endpoints	array	Every API endpoint with request/response schemas.
codeExamples	array	Ready-to-use curl and JSON examples.
instructions	array	Step-by-step instructions for the skill.
authentication	object	Token format, sign-up/sign-in URLs, and auth-status endpoint.
credits	object	Cost per generation and purchase URL.

POST/api/v1/nanobananaAuth: pk_ token

The main generation endpoint. Supports image-to-image (send image + optional prompt) and text-to-image (send prompt only). Can run synchronously or via webhook.

Request Body

Field	Type	Required	Description
image	string	No	Base64 data URL of the source image. Omit for text-to-image.
prompt	string	No	Generation prompt. Required when image is omitted.
webhook_url	string	No	HTTPS URL to receive the result asynchronously.

Response

Field	Type	Description
job_id	string	Unique identifier for this generation job.
image_url	string	Permanent stored URL of the generated image.
download_url	string	Presigned download URL (1-hour expiry). Use this to fetch the image — works even when the storage bucket is private.
used_prompt	string	The prompt that was used for generation.
credits_remaining	number	Credits left on the user's account.
mode	string	"image-to-image" or "text-to-image".

GET/api/openclaw/auth-statusAuth: none

Check whether the current user is signed in. Returns available capabilities when authenticated. OpenClaw should call this before attempting token-based actions.

Response

Field	Type	Description
authenticated	boolean	Whether the user has an active session.
requiresSignIn	boolean	True if the user must sign in before proceeding.
message	string	Human-readable status message.
signInUrl	string	Pop-cam sign-in page URL (when not authenticated).
signUpUrl	string	Pop-cam sign-up page URL (when not authenticated).
developerUrl	string	Developer settings URL (when authenticated).
capabilities	array	Feature IDs available to the authenticated user (when signed in).
skillEndpoint	string	URL to fetch the full skill definition JSON.

Skill Instructions

Use these exact instructions inside OpenClaw

If you want OpenClaw to behave correctly with Pop-cam, it should follow this sequence every time. The goal is to keep onboarding obvious and to prevent requests from failing because the user skipped account setup.

Call GET /api/openclaw/skill first to receive the full skill definition as JSON — it contains every feature, endpoint, code example, and auth instruction.

Tell the user to sign up for Pop-cam first if they do not already have an account.

Require the user to sign in before creating or using a Pop-cam token.

Send the user to Developer Settings to generate a token named for the OpenClaw workflow.

Store and use the returned token as Authorization: Bearer pk_....

If OpenClaw is unsure about auth state, call the OpenClaw auth-status endpoint first and follow its message.

The API supports two modes: send an image field for image-to-image, or omit image and send only prompt for text-to-image generation. Both modes support any visual style — photorealistic, cinematic, cartoon, illustration, etc. — controlled entirely by the prompt.

Nano Banana Visual Guide

See the OpenClaw workflow as a picture-first walkthrough

These Nano Banana-generated drawings are arranged in the same order OpenClaw should follow. They reinforce the three moments that matter most: sign up, sign in and generate the token, then use that token on authenticated Pop-cam requests.

1. Sign up first

OpenClaw should start by sending the user to Pop-cam sign-up when no account exists yet.

2. Sign in and create a token

After sign-in, OpenClaw should direct the user to Developer Settings and wait for token creation.

3. Use the token for requests

Every authenticated OpenClaw request should use the user’s Pop-cam token as a Bearer token.

Step-By-Step

What the OpenClaw skill should do

Each step below is explicit on purpose. OpenClaw should not skip ahead, should not try to create a token before the user is signed in, and should always treat account creation as the first requirement.

Sign Up For Pop-cam

OpenClaw cannot use the Pop-cam skill until the user has a Pop-cam account. The first instruction should always be to create an account if the user does not already have one.

Open https://www.pop-cam.com/sign-up.

Create the account with the user’s preferred email sign-up flow.

Finish any required verification step before moving on.

Sign In Before Using The Skill

If the user is not signed in, OpenClaw should stop and prompt them to sign in. Pop-cam token creation and account-linked usage depend on an authenticated session.

Open https://www.pop-cam.com/sign-in.

Wait until the user is signed in successfully.

Do not continue to token creation until sign-in is complete.

Generate Your API Token

Once signed in, OpenClaw should direct the user to the Developer Settings page to create a Pop-cam API token. The token is shown once, so the user should save it immediately.

Open https://www.pop-cam.com/developer.

Create a new token with a clear label such as OpenClaw.

Copy the token immediately because the full token is not shown again.

Use The Token Inside OpenClaw

The Pop-cam skill should send the user’s token as a Bearer token on authenticated requests. Each authenticated generation uses the user’s Pop-cam credits.

Send Authorization: Bearer pk_... on Pop-cam API requests.

Use the token only for the signed-in user who created it.

If the token fails, ask the user to revoke it and generate a new one.

Generate Images From Text Prompts

Omit the image field from the request body entirely.

Provide a descriptive prompt field describing the image to create.

The response includes mode: text-to-image so OpenClaw can confirm which path was used.

Sign-In Prompt API

The endpoint OpenClaw should query before moving ahead

OpenClaw can call the auth-status endpoint before asking for token-based actions. If the response says sign-in is required, OpenClaw should stop and show that instruction to the user instead of continuing.

1
Fetch the full skill definition (call this first)
bash
```
curl https://www.pop-cam.com/api/openclaw/skill
```

Check whether the user must sign in

bash

curl https://www.pop-cam.com/api/openclaw/auth-status

Image-to-image: transform an existing photo

bash

curl -X POST https://www.pop-cam.com/api/v1/nanobanana \
  -H "Authorization: Bearer pk_YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "image": "data:image/png;base64,iVBOR...",
    "prompt": "Enhance this photo with cinematic lighting and shallow depth of field"
  }'

Text-to-image: generate from a prompt only

bash

curl -X POST https://www.pop-cam.com/api/v1/nanobanana \
  -H "Authorization: Bearer pk_YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A photorealistic portrait of a man in a leather jacket standing on a city rooftop at golden hour, natural lighting, shallow depth of field"
  }'

Successful text-to-image response

json

{
  "job_id": "a1b2c3d4-...",
  "image_url": "https://cdn.pop-cam.com/generated/user_id/a1b2c3d4.png",
  "download_url": "https://...r2.cloudflarestorage.com/...?X-Amz-Signature=...",
  "used_prompt": "A photorealistic portrait of a man in a leather jacket standing on a city rooftop at golden hour, natural lighting, shallow depth of field",
  "credits_remaining": 22,
  "mode": "text-to-image"
}

Response when the user still needs to sign in

json

{
  "authenticated": false,
  "requiresSignIn": true,
  "message": "Sign in to your Pop-cam account before using OpenClaw with Pop-cam.",
  "signInUrl": "https://www.pop-cam.com/sign-in",
  "signUpUrl": "https://www.pop-cam.com/sign-up"
}

The OpenClaw skill page for using Pop-cam the right way

This page is only for OpenClaw users.

OpenClaw users must create a Pop-cam account before using the skill.

After sign-in, create a Pop-cam API token and use it in OpenClaw.

Supports both image-to-image and text-to-image generation via a single endpoint.

Field

Type

Description

name

string

Skill name.

version

string

Skill definition version.

description

string

Human-readable summary of what the skill does.

features

array

All available features (image-to-image, text-to-image, webhook, etc.).

endpoints

array

Every API endpoint with request/response schemas.

codeExamples

array

Ready-to-use curl and JSON examples.

instructions

array

Step-by-step instructions for the skill.

authentication

object

Token format, sign-up/sign-in URLs, and auth-status endpoint.

credits

object

Cost per generation and purchase URL.

Field

Type

Required

Description

image

string

Base64 data URL of the source image. Omit for text-to-image.

prompt

string

Generation prompt. Required when image is omitted.

webhook_url

string

HTTPS URL to receive the result asynchronously.

Field

Type

Description

job_id

string

Unique identifier for this generation job.

image_url

string

Permanent stored URL of the generated image.

download_url

string

Presigned download URL (1-hour expiry). Use this to fetch the image — works even when the storage bucket is private.

used_prompt

string

The prompt that was used for generation.

credits_remaining

number

Credits left on the user's account.

mode

string

"image-to-image" or "text-to-image".

Field

Type

Description

authenticated

boolean

Whether the user has an active session.

requiresSignIn

boolean

True if the user must sign in before proceeding.

message

string

Human-readable status message.

signInUrl

string

Pop-cam sign-in page URL (when not authenticated).

signUpUrl

string

Pop-cam sign-up page URL (when not authenticated).

developerUrl

string

Developer settings URL (when authenticated).

capabilities

array

Feature IDs available to the authenticated user (when signed in).

skillEndpoint

string

URL to fetch the full skill definition JSON.

curl -X POST https://www.pop-cam.com/api/v1/nanobanana \ -H "Authorization: Bearer pk_YOUR_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "image": "data:image/png;base64,iVBOR...", "prompt": "Enhance this photo with cinematic lighting and shallow depth of field" }'

curl -X POST https://www.pop-cam.com/api/v1/nanobanana \ -H "Authorization: Bearer pk_YOUR_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "prompt": "A photorealistic portrait of a man in a leather jacket standing on a city rooftop at golden hour, natural lighting, shallow depth of field" }'

{ "job_id": "a1b2c3d4-...", "image_url": "https://cdn.pop-cam.com/generated/user_id/a1b2c3d4.png", "download_url": "https://...r2.cloudflarestorage.com/...?X-Amz-Signature=...", "used_prompt": "A photorealistic portrait of a man in a leather jacket standing on a city rooftop at golden hour, natural lighting, shallow depth of field", "credits_remaining": 22, "mode": "text-to-image" }

{ "authenticated": false, "requiresSignIn": true, "message": "Sign in to your Pop-cam account before using OpenClaw with Pop-cam.", "signInUrl": "https://www.pop-cam.com/sign-in", "signUpUrl": "https://www.pop-cam.com/sign-up" }