AI Power Grid
Generate images cheaply (from $0.003 each) and run text models on AI Power Grid (AIPG), a decentralized GPU network with an OpenAI-compatible API and pay-per-use credit. Use when the user asks for an image and no other image provider is set up, mentions AIPG or AI Power Grid, or wants a low-cost pay-per-use model.
AI Power Grid (AIPG)
A network of independent GPU operators behind one API at https://api.aipowergrid.io/v1. It is
OpenAI-compatible and billed per use from the account's credit. This skill is community-written and not
an official AIPG document.
What it is good for
- Images. Several image models, from $0.003 per image, also for editing an existing image. This is the main reason to use it.
- Video. $0.02 per second, up to 26 seconds.
- Text. A few open models, cheap per token. For agent work with tool calls use
qwen38-flash-next-nvfp4.gpt-oss-120breturns tool calls too, but in tests on 2026-11-17 it sometimes reported a step as done without calling the tool.
Not covered here: audio and 3D. See the documentation at https://aipowergrid.io/docs.
Step 1: is there a key?
Check for the environment variable AIPG_API_KEY (a key starts with grid_), or a saved profile:
openclaw models auth list shows aipg: entries.
If there is none, do not try to make one and do not ask the user to paste a key into the chat. Tell the user:
- Open https://console.aipowergrid.io/?callbackUrl=%2Fdashboard%2Fapi-key and sign in (Google, GitHub or a wallet).
- Create an API key there.
- Store it without showing it to you, either
export AIPG_API_KEY=grid_...in the shell that starts OpenClaw, oropenclaw models auth paste-api-key --provider aipg.
An account whose sign-in wallet holds 10,000 AIPG or more on Base gets extra free credit each day; the tiers
are in the credit summary under free.holder_tiers. Signing in or paying with a wallet is not part of this
skill: the plugin's README describes it, and it is the user's decision and the user's command to run.
A new account may have no credit. Accounts with a verified Google sign-in get a small daily free allowance; otherwise the user adds credit at https://console.aipowergrid.io/dashboard/funding.
Step 2: prefer the plugin
Check: openclaw plugins inspect aipg. If it is installed, use it:
- Images: call
image_generatewithmodel: "aipg/z-image-turbo"(or anotheraipg/...image model). Sizes: 1024x1024, 1344x768, 768x1344, 1216x832, 832x1216. Up to 4 images per call. - Editing: pass one source image to
image_generatetogether with the instruction. - Video: call
video_generatewithmodel: "aipg/LTX-2.5 Distilled BF16"anddurationSeconds. - Text: the models appear as
aipg/<name>inopenclaw models list --refresh. aipg_balanceshows the credit (free). Call it before work that costs more than a few cents.aipg_councilasks several grid models the same question and says whether they agree. Use it when the user wants a second opinion.
If it is not installed, suggest it to the user (it needs their consent, so do not install it silently):
openclaw plugins install clawhub:openclaw-aipg-provider --accept-capabilities
Step 3: without the plugin, plain HTTP
Find out what is online and what it costs (both public, no key needed):
curl -s https://api.aipowergrid.io/v1/status/models # name, type (text/image/video/audio), count = workers online
curl -s https://api.aipowergrid.io/v1/pricing # price_book.models[].rates, e.g. per_image_usd
Use only models whose count is above 0. Send a non-default User-Agent: the API's edge refuses some default
client strings (Python's urllib gets HTTP 403).
Generate an image and save it:
curl -s https://api.aipowergrid.io/v1/images/generations \
-H "Authorization: Bearer $AIPG_API_KEY" -H "Content-Type: application/json" \
-d '{"model":"z-image-turbo","prompt":"a small red lighthouse at dusk, flat illustration","n":1,"size":"1024x1024"}'
The answer is {"data":[{"url":"https://..."}]} (sometimes b64_json instead of url). Download the URL to a
file and give the user the path. An image takes about 10 to 60 seconds.
Text, OpenAI chat format:
curl -s https://api.aipowergrid.io/v1/chat/completions \
-H "Authorization: Bearer $AIPG_API_KEY" -H "Content-Type: application/json" \
-d '{"model":"qwen38-flash-next-nvfp4","messages":[{"role":"user","content":"Hello"}],"max_tokens":1024}'
Keep max_tokens at 1024 or more: the models think before they answer, and with a small budget the answer
comes back empty but is still billed.
Credit left:
curl -s https://api.aipowergrid.io/v1/account/credits -H "Authorization: Bearer $AIPG_API_KEY"
total_spendable_usd is what can be spent now; free is today's allowance.
What the errors mean
| HTTP | Meaning | What to do |
|---|---|---|
| 401 | The key is wrong or was revoked | Ask the user to check the key; do not retry |
| 402 | Not enough credit | Tell the user; it is not a configuration error |
| 404 | The model is not online | Read /v1/status/models again and pick one that is |
| 413 | Request above 200,000 characters | Shorten or split the text |
| 429 | More than 30 requests in a minute | Wait a minute |
| 524 | The request ran longer than 120 seconds | Use "stream": true for long text answers |
Privacy
Whatever is sent is processed in plain text on a computer run by an independent operator. Do not send secrets, credentials or personal data. For images that means: keep such things out of the prompt.
Before the first paid request in a conversation, tell the user that it costs credit and roughly how much.
- Downloads
- Security audit
- Pass
- Last updated
- 1d ago
- Current version
- v0.4.2
- License
- MIT-0
