This source did not publish a separate summary. Review SKILL.md before using the skill.
SKILL.md
babbleBrush
babbleBrush is a voice-first image editor. Humans can use it directly to edit, view, manage images and agents can use it alongside their human or on behalf of the user to do the same programatically.
The units of work are canvases and versions, you create or use an existing canvas and edit the image inside of it as a canvas version, ie, the image edit. You can also upload images in the canvas to edit them, to use as a reference, or mixing them in the edit.
A Canvas is a workspace for image generation & editing.
A canvas contains multiple Canvas versions representing the edit history.
The latest completed canvas version is the current image of any given canvas.
Canvas version
A canvas version represents an image edit inside a canvas, it's the main unit of work.
It contains information all the information about the eidt, like the prompt, image references uploaded(the limit depends on the model), the source canvas version the edit branched from, the status, etc
Authentication
All requests require an API key sent in the Authorization header.
Authorization: Bearer bb_...
You can create API tokens from the settings page at https://babblebrush.com/app/settings.
Once you have the API key, store it locally and securely.
This is the main editing endpoint. It creates a new version and starts AI processing.
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "add a red ball in the center"}' \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions"
Optional parameters:
source_version_id - Edit from a specific version instead of the current version
model - Override the model for this specific edit (e.g., gemini/gemini-3-pro-image-preview)
How image inputs are chosen for this edit:
The new pending edit version points to a source_version.
source_version is source_version_id if provided, otherwise the canvas current version.
Image inputs for the provider come from that source_version.
If you want extra context images for the next edit, upload them as reference images to that source_version before enqueueing.
Max images are model-specific. Use GET /api/v1/provider-models and read max_imgs_per_prompt (total images = source image + reference images).
Reference images workflow:
Upload reference image(s) to the version you want to use as source.
Enqueue a new edit with that version as source_version_id.
# 1) Upload image to version 5 (source for next edit)
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/reference.png" \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions/5/images"
# 2) Enqueue edit from version 5 so uploaded image is used
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"make it cinematic", "source_version_id": 5}' \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions"
Example: add 3 reference images (extra context)
# Check model limits first (max_imgs_per_prompt)
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/provider-models"
# Upload 3 extra reference images to source version 5
curl -X POST -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/ref-1.png" \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions/5/images"
curl -X POST -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/ref-2.png" \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions/5/images"
curl -X POST -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/ref-3.png" \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions/5/images"
# Enqueue edit from that source version
curl -X POST -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt":"blend these references into one cohesive scene", "source_version_id": 5}' \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions"
Note: this 3-reference example needs a model that supports at least 4 total images (1 source + 3 references).
Model Override Example:
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "add a red ball", "model": "xai/grok-imagine-image"}' \
"https://babblebrush.com/api/v1/canvases/01234567-89ab-cdef-0123-456789abcdef/versions"
Model Resolution Order:
Per-edit model parameter (if provided)
Canvas default_model (if set)
Default model for the provider
Response (202 Accepted):
{
"version": {
"id": 6,
"version_number": 6,
"status": "pending",
"prompt": "add a red ball in the center",
"model": "gemini/gemini-2.5-flash-image",
"fail_reason": null,
"source_version_id": 5,
"is_current": false,
"image_url": null,
"reference_images_count": 0,
"created_at": "2025-10-11T13:00:00.000Z"
},
"remaining_credits": 9
}
Important: The edit is processed asynchronously. Poll the version endpoint or canvas to check when it completes.
Important: Cancel is best-effort. If provider request already started, provider-side compute may continue.
Common Workflows
1. Create canvas, enqueue edit, poll to completion
# 1) Create a blank canvas
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"title": "My Project"}' \
"https://babblebrush.com/api/v1/canvases/blank"
# 2) Copy `public_token` from response and set it
CANVAS_PUBLIC_TOKEN="<paste_public_token_here>"
# 3) Enqueue first edit
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "draw a mountain landscape with snow"}' \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
# 4) Copy `version.id` from response and set it
VERSION_ID="<paste_version_id_here>"
# 5) Poll this version (repeat until status is completed/failed/cancelled)
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/$VERSION_ID"
# 6) Read latest canvas state
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN"
2. Add reference images, then edit from that source version
# 1) Read canvas details
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN"
# 2) Copy `current_version_id` from response and set it
SOURCE_VERSION_ID="<paste_current_version_id_here>"
# 3) Upload extra context images to that source version
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/ref-1.png" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/$SOURCE_VERSION_ID/images"
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-F "image=@/path/to/ref-2.png" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/$SOURCE_VERSION_ID/images"
# 4) Enqueue edit from the same source version
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d "{\"prompt\":\"blend the scene using the references\", \"source_version_id\": \"$SOURCE_VERSION_ID\"}" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
3. Branch from an older version (with optional model override)
# Branch A from version 2 using default canvas model
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "try watercolor style", "source_version_id": 2}' \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
# Branch B from same source with explicit model override
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "try watercolor style", "source_version_id": 2, "model": "gemini/gemini-3-pro-image-preview"}' \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
4. Promote or cancel an edit
# Promote version 3 to current ("time travel")
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/3/promote"
# Cancel version 8 if still pending/processing (best-effort)
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/8/cancel"
5. Start multiple edits from the same source (different models)
# 1) Read canvas details
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN"
# 2) Copy `current_version_id` from response and set it
SOURCE_VERSION_ID="<paste_current_version_id_here>"
# 3) Create edit A from the same source using Gemini Pro
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d "{\"prompt\":\"cinematic dramatic lighting\", \"source_version_id\": \"$SOURCE_VERSION_ID\", \"model\": \"gemini/gemini-3-pro-image-preview\"}" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
# 4) Copy `version.id` from edit A response and set it
VERSION_A="<paste_version_a_id_here>"
# 5) Create edit B from the same source using xAI
curl -X POST \
-H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
-H "Content-Type: application/json" \
-d "{\"prompt\":\"cinematic dramatic lighting\", \"source_version_id\": \"$SOURCE_VERSION_ID\", \"model\": \"xai/grok-imagine-image\"}" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions"
# 6) Copy `version.id` from edit B response and set it
VERSION_B="<paste_version_b_id_here>"
# 7) Poll each edit (repeat each request until terminal status)
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/$VERSION_A"
curl -H "Authorization: Bearer $BABBLEBRUSH_API_KEY" \
"https://babblebrush.com/api/v1/canvases/$CANVAS_PUBLIC_TOKEN/versions/$VERSION_B"
Error Behavior
Status
Meaning
401
Missing or invalid API key
402
Insufficient credits
403
Forbidden
404
Resource not found
422
Invalid request (validation error)
500
Server error
Error response format:
{
"error": "Human-readable error message"
}
Rate Limits
Currently no rate limits are enforced, but please be respectful of the service.
Tips for Agents
Check access first - Before starting a workflow, verify the user has sufficient credits or platform_access_unlocked with provider keys configured
Poll for completion - Edits are async; poll GET /api/v1/canvases/:canvas_id/versions/:id until status is terminal (completed, failed, cancelled)
Use descriptive prompts - The AI responds better to detailed, clear prompts
Branch experiments - Use source_version_id to try different approaches without losing progress
Promote to revert - If an edit doesn't work out, promote an earlier version to go back
Inline Images in the chat - If the chat interface you're using allows(eg: Telegram), let's display the images inline instead of just posting the image link, unless the user asks for it.
