Integration guide

Use GLM-5.3 with Claude Code: Setup & API Guide

Install Claude Code, create a OneInfer API key, then point ~/.claude/settings.json at OneInfer's Anthropic-compatible endpoint with GLM-5.3 pinned as the model. Verify the connection with a small task before running it on a repository.

Create account & add creditsGet API Key

Prerequisites

  • Node.js 18 or newer
  • A OneInfer account with API credits
  • macOS: use nvm to install Node.js if you hit npm permission issues
  • Windows: install Git for Windows for a working bash/terminal environment

Step 1: Install Claude Code

  1. 1Install the CLI globally with npm.
  2. 2Move into your project directory.
  3. 3Run claude to launch it for the first time.
  4. 4Pick any terminal theme when prompted (Dark, Light, colorblind-friendly, or ANSI-only) — change it later with /theme.
Example
npm install -g @anthropic-ai/claude-code

cd your-project-directory

claude

Step 2: Create a OneInfer API key

  1. 1Sign in or create a OneInfer account.
  2. 2Open API Keys in the console and create a new key.
  3. 3Copy it once — OneInfer will not show it again — and store it in a secrets manager or environment variable, not in project files.
Get API Key

Step 3: Point Claude Code at GLM-5.3

Edit ~/.claude/settings.json and set the OneInfer endpoint, your key, and GLM-5.3 as the pinned model. Commands that change environment variables often print no output — that is normal when there is no error.

Example
{
  "$schema": "https://json.schemastore.org/claude-code-settings.json",
  "permissions": {
    "defaultMode": "acceptEdits"
  },
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.oneinfer.ai/v1/ula",
    "ANTHROPIC_AUTH_TOKEN": "<ONEINFER_API_KEY>",
    "ANTHROPIC_API_KEY": "<ONEINFER_API_KEY>",
    "ANTHROPIC_MODEL": "glm-5.3"
  }
}

Step 4: Verify the connection

  1. 1Open a new terminal so Claude Code reloads the environment, then run claude again in your project directory.
  2. 2If prompted "Do you want to use this API key?", select Yes and grant folder trust.
  3. 3Ask a short, deterministic task such as explaining one local function.
  4. 4Run /status and confirm it reports glm-5.3 before sending a large repository or document.
Example
claude
/status

Ready to test the workflow?

Create account & add credits

Model switching and defaults

You can pin per-tier defaults (haiku/sonnet/opus slots) instead of hardcoding one model, but for a GLM-5.3 evaluation it is simplest to keep ANTHROPIC_MODEL set explicitly so every request uses it.

Example
{
  "env": {
    "ANTHROPIC_MODEL": "glm-5.3"
  }
}

Troubleshooting

SymptomCheck
401 or 403Confirm ANTHROPIC_AUTH_TOKEN/ANTHROPIC_API_KEY and that the key is active.
Model not foundUse the exact model identifier exposed by the OneInfer catalog (glm-5.3).
Settings not applyingClose all Claude Code windows, open a new terminal, and run claude again.
Still brokenDelete ~/.claude/settings.json, validate the JSON (no trailing commas), and reconfigure — shell-level env vars can override settings.json.
High usageReduce supplied context and cap output for routine tasks.

Credential safety

Never paste an API key into a prompt, commit it to a repository, or expose it in browser-delivered code.

Frequently asked questions

Can I try GLM-5.3 before integrating it?

Use the OneInfer GLM-5.3 launcher to open a prepared prompt in the authenticated playground. Availability is checked against the current model catalog.

How should I treat benchmark claims?

Check the provenance label and harness version. Vendor-reported and independently verified results are deliberately shown as different evidence classes.

How do I check or update my Claude Code version?

Run "claude --version" to see the current version and "claude update" to upgrade to the latest release.

Put GLM-5.3 to work

Fund a controlled evaluation, start with a prepared prompt, and measure quality and cost on your own workload.

Create account & add creditsRun this with GLM-5.3