Quickstart
You'll have a working agent run in about five minutes. By the end of this page you'll have:
- Signed in to the Catentio portal with your Huudis identity.
- Invoked an agent and watched the run progress.
- Inspected the run's cost, artifacts, and event timeline.
If you'd rather read about concepts first, jump to Concepts. If you'd rather install the CLI or SDK first, see Installation.
Prerequisites
You'll need:
- A Huudis account whose
submatches the deployment'sHUUDIS_ALLOWED_USER_ID. Catentio v1 is single-user; the portal will reject any other Huudis identity at callback. - About five minutes.
- Either a browser (to use the portal) or the CLI installed (covered in Installation).
You do not need to configure tools, skills, or integrations to take an agent for a test drive. The runtime ships with sensible defaults and the cimi agent is callable out of the box.
1. Sign in
Head to your Catentio portal — for the canonical instance, that's catent.io. Click Log in.
The portal redirects to Huudis for authentication. You enter your Huudis email and password (or use Google / Apple, if your Huudis instance has those providers configured). Huudis redirects you back to /callback, where Catentio:
- Exchanges the one-time code for a Huudis access + refresh token.
- Verifies your
subclaim againstHUUDIS_ALLOWED_USER_ID. - If allowed, HMAC-signs a
catentio_sessioncookie and lands you on/dashboard.
If you see forbidden_user, you're not the configured single-user. See the Authentication overview.
Workspaces are how Catentio isolates customer-facing runtimes in the multi-tenant SaaS topology. In v1 (single-user, internal mode), you'll see exactly one workspace bound to your identity.
2. Invoke an agent
In the dashboard sidebar, click Agents. You'll see the agent roster — cimi, tora, sai, ren, fumi, iro, and the rest. Click any agent (start with cimi) to open the detail page.
Click Invoke. A dialog opens with one text field:
Message: Summarise the project board in three bullets.
Type your prompt and click Invoke. The portal POSTs to /api/v1/agents/cimi/invoke and immediately shows you the new run with status running.
The same call from the CLI:
catentio-saas agents invoke cimi --message "Summarise the project board in three bullets."
Or from Node:
import { CatentioSaasClient } from '@forjio/catentio-saas-node';
const catentio = new CatentioSaasClient({ apiKey: process.env.CATENTIO_API_KEY });
const run = await catentio.agents.invoke('cimi', {
message: 'Summarise the project board in three bullets.',
});
console.log('Run:', run.id);
Or from Python:
from catentio_saas import CatentioClient
with CatentioClient(api_key=os.environ["CATENTIO_API_KEY"]) as catentio:
run = catentio.agents.invoke("cimi", {"message": "Summarise the project board in three bullets."})
print("Run:", run["id"])
Or from Go:
import catentio "github.com/hachimi-cat/saas-catentio/sdk/go"
client, _ := catentio.NewClient(catentio.ClientOptions{
APIKey: os.Getenv("CATENTIO_API_KEY"),
})
run, _ := client.Agents.Invoke(ctx, "cimi", catentio.AgentInvokeInput{
Message: "Summarise the project board in three bullets.",
})
fmt.Println("Run:", run.ID)
3. Watch the run
Either way you invoked, the run shows up in Dashboard → Runs with a live status badge. Click into it and you'll see:
- The agent's streaming output (rendered as markdown).
- A subtask tree, if the agent decomposed the work via the task runner.
- Token + dollar cost (updated live; final on completion).
- An event timeline pulled from the runtime's outbox.
- Any artifacts the run produced (files, screenshots, JSON payloads).
When the status flips to succeeded, the agent's response is the run's final output.
Cost tracking is per-run, per-agent, per-model. If you want the workspace-level rollup, head to Dashboard → Cost. The dashboard sums across runs by day, week, and month and breaks the spend down by agent.
4. Verify it worked
The run's status should now be succeeded. Check from the portal at Dashboard → Runs, or via API:
catentio-saas runs list --limit 1
You did it — you've completed your first Catentio agent run.
What's next
- Installation — install the CLI or an SDK for repeated use.
- Concepts — the model behind agents, runs, projects, and skills.
- Portal → Agents — managing the agent roster.
- API reference — every endpoint, every parameter.