Integrations

Run your agent on ALEXIS Fusion

ALEXIS Fusion speaks the wires coding agents already use, so the tools you have keep working. Point one at ALEXIS and it runs on the fusion model instead of a single vendor. Your existing setup is never modified: every recipe here is additive, and the plain command always goes back to whatever you had.

The three values

Every integration comes down to the same three settings. Wherever a tool lets you set an endpoint, a key, and a model, use these.

Endpoint: https://api.alexisengine.com/v1
API key: Create one in the dashboard
Model: alexis-fusion

Create an API key

Claude Code

Verified live

Runs natively. One command installs the launcher and stores your key; after that one word starts Claude Code on ALEXIS Fusion.

Once

export ALEXIS_API_KEY="your key"
curl -fsSL https://dashboard.cgai.co/alexis-setup.sh | sh

Anytime

alexis launch claude

Nothing is written into Claude Code's own configuration, so plain claude stays exactly as you had it.

OpenAI Codex CLI

Verified live

Runs natively. The installer adds a separate profile that Codex reads only when you ask for it.

Once

export ALEXIS_API_KEY="your key"
curl -fsSL https://dashboard.cgai.co/codex-setup.sh | sh

Anytime

alexis launch codex

Plain codex keeps your normal account and defaults, login and all.

Hermes

Verified live

Supports custom endpoints natively. Point it at ALEXIS once, then run Hermes the way you always do.

Hermes app: Settings -> Providers -> Custom Endpoints

Settings -> Providers -> Custom Endpoints -> New endpoint

  Name           ALEXIS Fusion
  Provider ID    custom
  Endpoint URL   https://api.alexisengine.com/v1
  Default Model  alexis-fusion
  Context        Auto
  API Key        Paste your ALEXIS Fusion API key here

  [x] Use for new chats
  [x] Discover models

Click Test, then Save. The endpoint row shows Active once Hermes is
running on it; Use switches an existing endpoint back on.

Or the config file

cat >> ~/.hermes/config.yaml <<'EOF'
model:
  provider: custom
  base_url: https://api.alexisengine.com/v1
  model: alexis-fusion
  api_key: your key
EOF

Anytime

hermes

Or one shot, no setup at all

hermes --provider custom \
  --base-url https://api.alexisengine.com/v1 \
  --model alexis-fusion

Use the app screen or the config file, not both. If the config file already has a model section, edit that section instead of appending a second one: a duplicate model key makes Hermes unpredictable.

OpenClaw, OpenCode, GitHub Copilot CLI, and others

OpenAI-compatible

Any tool that lets you point at a custom OpenAI-compatible endpoint works with the three values above. Put the endpoint in the base URL field, your key in the API key field, and alexis-fusion in the model field.

Typical environment form

export OPENAI_BASE_URL="https://api.alexisengine.com/v1"
export OPENAI_API_KEY="your key"
export OPENAI_MODEL="alexis-fusion"

A tool that expects a specific vendor model id almost always has a custom-model or model-override field. Set it to alexis-fusion.

Not seeing your tool?

If it can talk to an OpenAI-compatible or Anthropic-compatible endpoint, it can talk to ALEXIS Fusion. Tell us what you use and we will confirm the recipe.

Let’s talk

Bring us your hardest workload.

Tell us what you are running today and where cost or accuracy hurts. We reply within one business day.

Prefer to start building? Create an API key.