CLI Quickstart¶
This walkthrough uses the current Go CLI command set and output shape.
Step 1: Install¶
Step 2: Configure¶
runagents config set endpoint https://your-workspace.try.runagents.io
runagents config set api-key ra_ws_your_workspace_key
runagents config set namespace default
runagents config get
Step 3: Seed Starter Resources¶
This creates starter resources (echo-tool, playground-llm) if missing.
Step 4: Create Agent File¶
Create agent.py:
Step 5: Deploy¶
runagents deploy \
--name hello-world \
--file agent.py \
--tool echo-tool \
--model openai/gpt-4o-mini \
--policy echo-read-policy
Notes:
runagents deploymaps toPOST /api/deploy.- You can bind policies and identity providers directly in the CLI with
--policyand--identity-provider.
Step 6: Verify Agent¶
Step 7: Inspect Runs¶
Step 8: Approvals (When Triggered)¶
If a policy rule returns approval_required, use:
runagents approvals list
runagents approvals approve <request-id> --scope once
# or approve the current run for one hour
runagents approvals approve <request-id> --scope window --duration 1h
# or
runagents approvals reject <request-id>
Production-oriented path: deploy from the catalog¶
The quickest first-run path is still Hello World, but many teams want to validate a real workflow next.
A strong next step is deploying a catalog agent such as the Google Workspace assistant. That path is better when you want to test:
- delegated-user OAuth
- policy-controlled writes
- approval-required operations
- pause and resume behavior across real tools
Example:
runagents catalog list --search google
runagents catalog show google-workspace-assistant-agent
runagents catalog deploy google-workspace-assistant-agent \
--name google-workspace-assistant-agent \
--tool email \
--tool calendar \
--tool drive \
--tool docs \
--tool sheets \
--tool tasks \
--tool keep \
--policy workspace-write-approval \
--identity-provider google-oidc
See Agent Catalog for the recommended production-style path.
Useful Commands¶
runagents tools list
runagents tools get echo-tool
runagents models list
runagents context export -o json
runagents identity-providers list
runagents policies list