KoKonna + OpenClaw: How to make an ultimate custom e ink display frame

OpenClaw runs on a separate device. The KoKonna frame remains a paper-like display for still images.

Quick answer: OpenClaw does not run inside a KoKonna frame. It runs on a computer or another supported device, where it can receive requests through its Control UI or a connected chat channel. A community-published kokonna-frame skill can then upload a selected image to a configured KoKonna frame, check device information and target more than one frame.

The connection has two layers. OpenClaw handles the conversation, tools and optional automation. KoKonna displays the finished image. The frame itself does not become a microphone, speaker or general-purpose AI assistant.

What the KoKonna and OpenClaw connection actually does

The four parts of the workflow have different jobs:

You do not speak to the wall frame. You send a request through OpenClaw's own interface or a connected service, and the configured workflow passes the finished image to KoKonna.

What the community KoKonna Frame skill can do

At the time of review, ClawHub listed Kokonna Frame as a community-published skill by guooscar. Its documentation covers these functions:

· Upload a local image to one named KoKonna frame.

· Crop and resize the image to the frame's detected display dimensions.

· Add a custom filename to the uploaded image.

· Query documented device information, including battery level, online status and firmware version.

· Send the same image to every frame listed in the local configuration.

OpenClaw coordinates the external workflow; KoKonna receives and displays the finished still image. A scheduled composition remains static on the E-Ink display until another image is selected.

The skill is an upload and device-control bridge. It does not generate artwork by itself. If you ask OpenClaw for a new image, an enabled model or image-generation tool must create the file first. The KoKonna skill then handles the final upload.

Where OpenClaw Runs—and What the Frame Displays

OpenClaw runs separately and does not change KoKonna's hardware. The frame does not gain a built-in microphone, speaker, backlight or video playback, and it does not become a continuously updating live dashboard.

The following functions depend on OpenClaw or separate tools rather than the KoKonna skill:

· AI image generation requires a configured model or image tool.

· Weather, news and calendar data require their own OpenClaw tools or integrations.

· Scheduled image changes require an OpenClaw automation that you create and maintain.

· Voice input, when available, comes from a phone, computer or supported OpenClaw channel rather than the frame.

KoKonna remains a quiet, reflective art display while the active AI work happens on the external OpenClaw device and its connected services.

Before you start

This is a technical integration rather than a standard KoKonna App feature. Prepare the following before installing anything:

· A KoKonna frame that has completed its normal setup and is online.

· A computer or compatible device running the current version of OpenClaw.

· The device API key shown on the KoKonna Frame settings page.

· Python 3.7 or later, plus the requests, Pillow and PyYAML packages.

· A test image that does not contain sensitive or private information.

The public ClawHub listing identifies this as a community package rather than a standard KoKonna App feature or an official OpenClaw skill. Review the current package files, version history and security information before installing it. Keep the device API key private and use only the current KoKonna OpenAPI host.

How to set up the OpenClaw E-Ink frame workflow

1. Install and start OpenClaw

Follow the current OpenClaw getting-started guide. The official installer for macOS, Linux and WSL2 is:

curl -fsSL https://openclaw.ai/install.sh | bash

Complete onboarding, start the Gateway and confirm that you can send a message through the Control UI before adding the KoKonna skill.

2. Install the community KoKonna Frame skill

OpenClaw's current ClawHub documentation uses an owner-qualified package reference. For the community package published by guooscar, use:

openclaw skills install @guooscar/kokonna-frame

Check the current KoKonna Frame skill listing before installation. The package owner, version and files may change after this article is published.

3. Install the Python dependencies

The current skill uses requests for API calls, Pillow for image processing and PyYAML for its configuration file:

python3 -m pip install requests Pillow PyYAML

Use a dedicated Python environment if that is already part of your OpenClaw setup.

4. Add your frame configuration

The current skill looks for a configuration file at:

~/.openclaw/skills/kokonna-frame/config.yaml

The documented structure is:

api_base_url: "https://api.kokonna.art/openapi"
devices:
living_room: "your-api-key-here"
bedroom: "your-second-api-key-here"

Use short device names that are easy to recognize. Do not publish this file, commit it to a public repository or paste its API keys into a shared chat.

5. Test one frame before using automation

From the installed skill directory, query a single device first:

python3 scripts/device_info.py --device living_room

If the device responds correctly, send a non-sensitive test image:

python3 scripts/upload.py \
--device living_room \
--image /path/to/test-image.jpg \
--name "Test Image"

Confirm the target frame before using --device all. That option sends the same image to every device in the configuration.

6. Use OpenClaw through your chosen interface

Once the skill is available to the agent, you can make requests such as:

· "Send this image to my living room frame."

· "Check the battery level of the bedroom frame."

· "Upload the selected artwork to all configured frames."

These requests go to OpenClaw through its Control UI or a channel you have separately configured. They do not go directly to the frame.

How scheduled artwork can work

OpenClaw Automations can schedule a workflow that obtains data, creates or selects a still image and then calls the KoKonna upload skill. The Gateway must remain running for scheduled jobs to execute, and these steps are not included automatically just because the frame skill is installed.

Before running a daily schedule, test the complete workflow manually. Check the image that will be uploaded, confirm the target frame and choose an update frequency that makes sense for an E-Ink display. A failed image-generation step should not trigger an upload of an empty or unintended file.

Image cropping and display fit

The community skill may crop and resize an image before upload. KoKonna's official OpenAPI also states that uploaded images are automatically cropped to the device display area. For important artwork, prepare the source image for the target frame ratio before uploading.

Leave enough space around faces, text and signatures, and test landscape and portrait compositions separately. The KoKonna frame displays still images rather than video or continuously updating content. Use the KoKonna App for the standard upload and creation workflow when automation is unnecessary.

Privacy and security checks

This workflow is not fully local. Selected images and device-status requests are sent to the configured KoKonna API.

Use a few basic safeguards:

· Keep frame API keys in the local configuration file and restrict access to it.

· Verify the API host before adding credentials.

· Review every private photo before allowing an agent to upload it.

· Confirm the target device list before any all-frame action.

· Recheck the community package after updates instead of assuming a new version behaves exactly like the previous one.

OpenClaw also advises users to treat inbound messages as untrusted input. If you expose the agent through a chat channel, use OpenClaw's pairing and access controls so an unknown sender cannot trigger frame actions.

KoKonna App or OpenClaw: which should you use?

For most owners, the KoKonna App remains the simplest choice. OpenClaw is better suited to developers and experienced users who want to experiment with agent-driven image workflows.

The intelligence runs elsewhere; KoKonna remains the paper-like display at the end of the workflow.

Frequently asked questions

Does OpenClaw run on or listen through the KoKonna frame?

No. OpenClaw runs on a separate supported device. Requests are made through OpenClaw's Control UI or a separately configured channel, not directly through the frame.

Is the KoKonna Frame skill official?

ClawHub currently lists it as a community skill published by guooscar. Review the current listing before installation and do not assume it has the same support status as the KoKonna App.

Does the skill create AI art?

No. It uploads an image file and can query device information. OpenClaw needs a separate model or tool to generate new artwork.

Will the skill crop my image?

The current community skill may crop and resize the source image, and KoKonna's official OpenAPI also states that uploads are automatically cropped to the device display area. Test compositions with important content near the edges.

Can OpenClaw change the frame automatically every day?

Yes, if you build and test an OpenClaw automation that obtains or creates an image and then calls the upload step. The Gateway must be running for scheduled jobs to fire.

Start with a Simple OpenClaw E-Ink Workflow

OpenClaw can coordinate a request and pass a finished still image to a KoKonna frame through a community skill. Keeping each part of the workflow separate makes the setup easier to test and avoids implying hardware features the frame does not have.

Start with one frame and one non-sensitive test image. Once the upload works reliably, add generation tools, channels or scheduled automations one at a time. For the standard experience, browse KoKonna E-Ink frames and use the KoKonna App. For advanced setup details, consult the OpenClaw skills documentation and Automations guide.