Claude Code

Who it’s for

Developers who want to use Claude Code with a GlideflowAI model through the Anthropic-compatible protocol.

Prerequisites

Claude Code installed
A GlideflowAI API key

Get your API key

Open the GlideflowAI dashboard, create a key, and keep it private.

Base URL

Use https://api.glideflowai.com without /v1.

Model ID

Start with glm-5.2. Copy other exact IDs from the model catalog.

Minimal working config

export ANTHROPIC_BASE_URL="https://api.glideflowai.com"
export ANTHROPIC_AUTH_TOKEN="sk-your-key"
export ANTHROPIC_MODEL="glm-5.2"
claude

You can also persist the same values in ~/.claude/settings.json:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.glideflowai.com",
    "ANTHROPIC_AUTH_TOKEN": "sk-your-key",
    "ANTHROPIC_MODEL": "glm-5.2"
  }
}

Test it

Run claude, then ask it to answer with OK. You can also test the same key and model directly:

curl https://api.glideflowai.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{"model":"glm-5.2","messages":[{"role":"user","content":"Say OK"}],"max_tokens":512}'

Common errors

Do not append /v1 to ANTHROPIC_BASE_URL.
If the response has no visible text, give the reasoning model at least 512 output tokens.
HTTP 503 with 无可用渠道 means the model ID is wrong or is not enabled for your key. Copy the exact ID from /models.

Notes / limitations

The Anthropic-compatible path accepts x-api-key; Claude Code uses the auth-token environment variable shown above. Model behavior differs by provider even though the tool connection works.