跳到主要内容

Subconscious Loop

A background task evaluation and execution system. On a periodic tick, it loads a list of user-defined and system tasks, reads the current state of your workspace, decides what to do about each one, and either acts autonomously or escalates to you for approval.

Think of it as the agent's idle thread: the part that keeps thinking after you've stopped typing.

How a tick works​

┌─────────────────────────────────────────────────────────┐
│                    Heartbeat                            │
│           (sleeps a few minutes between ticks)          │
└──────────────────────┬──────────────────────────────────┘
                       │
                       ▼
┌─────────────────────────────────────────────────────────┐
│                  Subconscious Engine                    │
│                                                         │
│  1. Load due tasks                                      │
│  2. Mark each one in-progress                           │
│  3. Build a situation report (memory + workspace)       │
│  4. Evaluate every task with the local model            │
│  5. Execute the decision (act / noop / escalate)        │
│  6. Write the outcome back to the activity log          │
└─────────────────────────────────────────────────────────┘

Each tick is independent. If a tick is still running when the next one starts, the new tick takes over and the old one's in-progress entries are marked cancelled. Ticks never stack.

Task types​

System tasks​

Seeded automatically when the engine starts. Cannot be deleted, only disabled:

  • Check connected skills for errors or disconnections
  • Review new memory updates for actionable items
  • Monitor system health

User tasks​

Anything you add manually from the UI:

  • "Check urgent emails" (read-only)
  • "Send daily summary to Slack" (write intent)
  • "Summarize Notion updates" (read-only)

Decisions​

For every due task, the local model returns one of three decisions:

DecisionMeaning
SkipNothing relevant right now
ActSomething relevant found, execute the task
EscalateNeeds deeper reasoning, hand off to the cloud agent

Two models, one loop​

StageWhere it runsWhy
Per-task evaluation (every tick)Local model (Ollama)Free, no rate limit, fine on-device
Tool-using execution (send, post, …)Cloud agentTools, larger context, retries on rate-limit
Analysis mode for escalated readsCloud agent (read-only)Deeper reasoning when the local model defers

Approval gate​

Approval is only required when the agent wants to take a write action that you didn't explicitly ask for.

Configuration​

The loop is configurable in the desktop app:

  • Enable / disable. Turn the entire background loop on or off.
  • Tick interval. How often a tick fires. Defaults to 5 minutes.
  • Inference. Whether the local model evaluates tasks each tick.
  • Context budget. How much of the workspace situation report can be passed in at once.

In the UI​

Lives under Intelligence → Subconscious:

  • Status bar. Task count, total ticks, last tick time, failure counter.
  • Active Tasks. System tasks and your own tasks.
  • Approval Needed. Amber cards for pending escalations.
  • Activity Log. Chronological feed of every task evaluation.
  • Run Now. Manually trigger a tick.

See also​