Agent model cards
PULSE does its AI work through a small set of agents. Each one has its own instructions, its own limits, and its own guardrails, but they all run on the same model server. A model card says, in plain language, what one agent does, what it reads, what it produces, who checks its work, and where it stops.
The cards
| Agent | What you meet it as |
|---|---|
| Diagnostic analyst | The written diagnosis from a Site Audit, and the metrics read-out |
| Market radar | Radar research, competitor channels, and guest spots |
| Strategy planner | Recalculating the route of a 30-day follow-up cycle |
| Channel content creator | Posts, emails, outreach drafts, and the shorter Studio pieces |
| Long-form studio | Whitepapers, blog posts, and the weekly audiobook |
| Partnership outreach | First-contact messages to podcasts, channels, and brands |
| Sales and voice agent | Cold-call scripts, live replies on a phone call, and churn watch |
| Customer 360 and journeys | Audience profiles, journey maps, content strategy, and roadmaps |
| PULSE assistant | The chat on every screen |
| Self-healing operations agent | Keeps PULSE’s own services running; the one agent that acts automatically, with an off switch |
What every agent has in common
The model. Every agent runs on the same base model:
Qwen3.6-35B-A3B, made by Qwen and released under the Apache-2.0 licence.
The deployment decides which model its inference server loads
(INFERENCE_MODEL); see Sovereign inference.
Runink domain adaptation for these agents is planned; this release uses the
base model.
Where it runs. On the deployment’s own inference server: your Runink Server, or the cloud account it runs in. PULSE calls no third-party AI service. Some agents read the public web or your connected accounts while they work, and PULSE talks to the services you connect (HubSpot, LinkedIn, Google Analytics and others) as described in Connections.
Text only. The agents read and write text. The current Server release does not include a vision model, so no agent looks at images or screenshots. Still-image generation is not available, and video needs the video tool the deployment configures. See Content pipeline.
Guardrails before the model. Before any agent is called, PULSE checks the request against a shared rule set. It refuses attempts to override the agent’s instructions, requests to reveal secrets or credentials, requests to reach internal network addresses, and hateful content. Each agent adds its own rules, listed on its card.
Outside text is data, not instructions. Web pages, search results, attached sources, lead and CRM records, inbound customer messages, and call transcripts reach an agent marked as material to read. Before they do, PULSE removes model control codes and hidden characters and takes out text that tries to give the agent instructions. Every agent is told never to follow instructions found in that material.
Value first. Every agent that writes for an audience follows the same brand rule: teach the reader something useful, mention your product at most once and only as an honest example, no hype words, no fake urgency, and end with a real question rather than a hard sell.
People decide. Agents draft, recommend, and score. (The self-healing operations agent is the exception for PULSE’s own infrastructure: it applies a fixed list of safe actions automatically, and an operator can turn that off.) Drafts go to the Approval review queue. A post is published only when a person approved that piece and an administrator armed the channel; rejecting a piece withholds its scheduled posts. The one automatic exception is the reply to an inbound WhatsApp or SMS message from a known contact: it is on by default, passes the same guardrails as every other reply (the message is read as data, injection attempts are refused, the reply is screened), and an administrator can turn it off. See Publishing controls. Some proposals also pass through the judging ladder, and the agent activity card shows every run as it happens.
Evaluation. No published evaluation scores yet, for any agent.