How to Use GLM-4.5V in OpenClaw
GLM-4.5V helps interpret screenshots, charts, documents, and visual problems. Its thinking mode suits tasks that need reasoning, while non-thinking mode suits straightforward reading and extraction.
Prompts to try with GLM-4.5V in OpenClaw
Copy an example into OpenClaw, or ask in your own words.
Every morning, use GLM-4.5V to summarize newly received report pages. Return the main findings with page references and flag any conclusion that depends on unreadable text.
Each weekday, use GLM-4.5V to extract the labels, units, and printed values from the latest dashboard chart image. Keep estimated values in a separate section.
Each evening, use GLM-4.5V to convert newly received scanned tables into rows for review. Preserve footnotes and flag unreadable cells without filling them in.
Every Friday, use GLM-4.5V to summarize the visible events in the supplied demonstration videos. Keep events in order and flag passages where the action is unclear.
After each scheduled interface capture, use GLM-4.5V to interpret the numbered screenshot against the task goal. Suggest one next element label to select and include the visible reason.
How to set up GLM-4.5V in OpenClaw
Pick the way that suits you. Each takes about two minutes.
Use the Spiral skill
- Send OpenClaw 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 Spiral's MCP server
- Add the server
On the machine that runs OpenClaw's Gateway:
openclaw mcp add spiral \ --url https://api.spiral.ai/mcp \ --transport streamable-http \ --auth oauth
- Sign in
In Settings › MCP, select Sign in on the Spiral server, sign in to Spiral and click Authorize. No API key.
openclaw mcp login spiral
- Reload and check
A running Gateway may keep its old tool list: restart it, or Save & Publish in the Control UI.
openclaw mcp reload openclaw mcp doctor spiral --probe
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 GLM-4.5V
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": "openrouter", "model": "z-ai/glm-4.5v", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: OpenClaw finds GLM-4.5V on Spiral, shows the price and runs it.
GLM-4.5V providers and prices
Spiral runs GLM-4.5V through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.009613 per run at the default inputs. $0.78 per 1M input tokens, $2.34 per 1M output tokens.
Settings you can ask for
Ask OpenClaw for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Seed | a whole number | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Max tokens | 1 to 16384 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Presence penalty | -2 to 2 | ||
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of Z.ai: GLM 4.5V is on its tool page: open its card above.
What can you do with GLM-4.5V in OpenClaw?
Turn recurring visual material into reviewable summaries with GLM-4.5V. An agent can examine incoming document pages, charts, and screenshots on a schedule and flag entries that need a person's judgment.
Turn a screenshot into a webpage draft
Ask GLM-4.5V to translate a reference screenshot into HTML and CSS, then review the rendered result against the reference.
Explain a chart
Extract the labels, units, and visible values, then ask which comparisons the chart supports. Keep estimates separate from printed numbers.
Extract a document table
Turn a table shown on a document page into rows for review. Ask it to retain headers, units, and footnotes and mark unreadable cells.
Review a research report
Summarize the report's argument and supporting charts. Request page references for each finding and separate the report's claims from the model's interpretation.
Explain a visual STEM problem
Ask for an explanation of a math or science problem that depends on a diagram. Have it identify the given values before calculating.
Summarize video events
Produce an ordered account of visible events in supplied video content. Ask for evidence for each event and flag uncertain transitions.
Choose the next screen action
Give the task goal and current screenshot, then request the next interaction. Numbered element labels can make the proposed target easier to check.
Locate an object in an image
Describe the target through its appearance and position, then request its bounding box. Review the proposed location against the image.
More on GLM-4.5V and OpenClaw
Overview
GLM-4.5V in other agents
More OpenClaw guides
Questions about GLM-4.5V in OpenClaw
Can OpenClaw use GLM-4.5V?
Yes. Add Spiral to OpenClaw as an MCP server, and OpenClaw can find and run GLM-4.5V like any built-in tool.
How much does GLM-4.5V cost in OpenClaw?
Up to $0.009613 per run at the default inputs. $0.78 per 1M input tokens, $2.34 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your OpenClaw plan is separate.
Do I need an OpenRouter account to use GLM-4.5V in OpenClaw?
No. Spiral runs GLM-4.5V on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up GLM-4.5V in OpenClaw?
About two minutes: add the server, then sign in, then reload and check.
Does OpenClaw wait while GLM-4.5V runs?
GLM-4.5V runs in the background. OpenClaw starts the run, checks on it and gives you the text, json when it's ready.
What is GLM-4.5V useful for?
Use it for questions that depend on visual evidence: interpreting documents, understanding screenshots, explaining diagrams, and analyzing video events. It also supports coding tasks grounded in visual input.
When should I use thinking mode?
Use thinking mode for tasks such as solving a diagram-based problem or explaining relationships across document content. For simple recognition or extraction, try non-thinking mode first.
How much context and output can GLM-4.5V handle?
GLM-4.5V supports a 64K-token context window. Z.ai documents a maximum output of 16K tokens.
Can GLM-4.5V summarize documents and extract chart data?
Yes. Z.ai lists document summarization, translation, and chart extraction among its uses. Ask for page references and preserve the chart's units so you can check the result.
Where does GLM-4.5V need extra checking?
Check object counts and claims about a person's identity. Z.ai also notes weaknesses in pure text question answering and says the model can overthink or repeat content.
Start using GLM-4.5V in OpenClaw
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into OpenClaw:
set up https://spiral.ai/SKILL.md


