Hyde Inference

Open weights.
Sovereign Frontier AI.

The leading open-source models, hosted in-region, behind an easy to configure OpenAI- and Anthropic-compatible endpoint.

No invite yet? Write to inference@hyde.ai

Artificial Analysis
“Kimi K3.1 Pro also achieves the highest score of any model on Terminal-Bench 4.0.”
Regional hubsUS · APAC · EU
Context1M tokens
Compatible withOpenAI + Anthropic APIs
ComplianceSOC 2 · HIPAA
Setup · about two minutes

Paste one config. Keep everything else.

Pick your tool, copy the config, done.

Choose your tool
Full guide ↗
echo "alias claude-hyde='CLAUDE_CONFIG_DIR=~/.claude-hyde claude'" >> ~/.zshrcmkdir -p ~/.claude-hyde && cat > ~/.claude-hyde/settings.json <<'EOF'{  "apiKeyHelper": "echo $HYDE_INFERENCE_API_KEY",  "model": "hyde/glm-5.3-flash",  "env": {    "ANTHROPIC_BASE_URL": "https://inference.hyde.tech",    "CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY": "1"  }}EOFecho 'export HYDE_INFERENCE_API_KEY="sk-PASTE-YOUR-KEY"' >> ~/.zshrcsource ~/.zshrcclaude-hyde
  1. Generate a key in the dashboard (it is shown once) and put it in place of sk-PASTE-YOUR-KEY.
  2. Paste the block into your terminal. It adds a claude-hyde command with its own profile, so your usual claude setup stays untouched. On bash, use ~/.bashrc.
  3. In claude-hyde, run /status. The base URL should read inference.hyde.tech.