> ## Documentation Index
> Fetch the complete documentation index at: https://www.rumus.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Z.AI

> Use your Z.AI key to run the GLM family in Rumus, including the Coding-specialized variant.

Z.AI provides the GLM model family — strong, cost-effective models with first-class tool calling and prompt caching. Rumus has a dedicated provider so you can use both the general endpoint and the Coding-specialized one.

## Before you start

You need:

* A Z.AI account at [z.ai](https://z.ai/) (or the international portal).
* An API key from your Z.AI dashboard.
* Sufficient credit on the Z.AI account.

## Add Z.AI in Rumus

<Steps>
  <Step title="Open the model settings">
    Go to **Settings → AI → Models** and click **Add Model**.
  </Step>

  <Step title="Pick the provider">
    Set **Provider** to **Z.AI**.
  </Step>

  <Step title="Paste your API key">
    Paste the key into **API Key**. Stored encrypted in your local [vault](/docs/account/vault).
  </Step>

  <Step title="(Optional) Custom base URL">
    Two endpoints are common:

    * **General**: `https://api.z.ai/api/paas/v4` (default — leave blank).
    * **Coding**: `https://api.z.ai/api/coding/paas/v4` — better routing for code-heavy workloads.

    Paste the Coding URL into **Base URL** if you want that variant.
  </Step>

  <Step title="Pick a model">
    Choose from the list (GLM-4.7, GLM-4.6, GLM-4.5, GLM-4.5 Air, GLM-4.5 Flash, etc.) or toggle **Enter custom ID** to type a model ID manually.
  </Step>

  <Step title="Capabilities">
    On the **Capabilities** tab:

    * **Tool Calling** — supported on GLM-4.5+ and 4.7.
    * **Vision** — supported on multimodal GLM versions.
    * **Prompt Cache** — supported on GLM-4.7; cache reads are \~50% of input price.
  </Step>

  <Step title="Save">
    The model appears in the picker under **Custom Models**.
  </Step>
</Steps>

## Recommended models

| Model             | Good for                                                    |
| ----------------- | ----------------------------------------------------------- |
| **GLM-4.7**       | Default — long context (128K), strong tools, prompt caching |
| **GLM-4.5 Air**   | Lightweight, faster, lower cost — great for autocomplete    |
| **GLM-4.5 Flash** | High-volume, latency-sensitive                              |

For the latest model details and limits, see the [Z.AI documentation](https://docs.z.ai/).

## Tips

* The default **temperature/top\_p** for Z.AI (0.7 / 0.9) differs from OpenAI's defaults. Rumus respects whatever you set per thread.
* **Coding endpoint vs general**: if you're mostly using the agent for code tasks, configure the Coding endpoint — routing and rate limits are tuned for it.
* **Prompt cache** kicks in automatically on supported models. Long, stable system prompts (skills, rules, long file context) benefit the most.

## Troubleshooting

<AccordionGroup>
  <Accordion title="401 Unauthorized">
    Recheck the API key. Note that Z.AI keys are not interchangeable between the China portal and the international portal.
  </Accordion>

  <Accordion title="Slow first responses">
    Streaming first-token latency varies by region. If you're far from the endpoint, switch to a closer region (Z.AI publishes regional URLs).
  </Accordion>

  <Accordion title="Model not in the list">
    Toggle **Enter custom ID** and paste the exact model ID (e.g. `glm-4.7`).
  </Accordion>
</AccordionGroup>

<Note>
  Hit a snag we didn't cover? Ask in the [Rumus community](https://www.rumus.ai/community).
</Note>

## Next steps

<CardGroup cols={2}>
  <Card title="Other providers" icon="grid" href="/docs/models/overview">
    Anthropic, OpenAI, Google, DeepSeek, Kimi, Ollama, OpenAI-compatible.
  </Card>

  <Card title="Built-in models" icon="cube" href="/docs/models/built-in">
    Use Rumus's bundled GLM-4.7 without managing a key yourself.
  </Card>
</CardGroup>