Lacesse Intelligence
The cognitive engine bridging the gap between non-linear ideation and linear execution.
Lacesse Intelligence is not a chatbot. It is a Hierarchical Reasoning Model (HRM) integrated directly into your project management workflow. Unlike traditional AI that exists in a silo, Lacesse Intelligence reads your codebase, understands your organization's vector memory, and actively manipulates the Lacesse One Project Board.
The Split-Mind Architecture
Complex software engineering requires two distinct modes of thinking: Exploratory (What could we build?) and Executive (How do we build it?). Lacesse Intelligence physically splits the interface to accommodate these modes.
Left Pane: The Planner
A node-based, branching chat interface. It handles ambiguity, allows for backtracking, and visualizes decision trees using our proprietary rendering engine.
Right Pane: The Builder
The Lacesse One Project Board. This is a linear, Kanban-style execution environment where abstract ideas are solidified into trackable tickets.
Branching Logic
Standard LLM interfaces are linear streams of text. This is inefficient for architectural planning. Lacesse Intelligence treats every response as a Node.
- Divergence: Users can click on any previous node to "fork" the conversation. The UI visually renders a new branch line.
- Visual Cues: Active branches glow with a Cyan aura. Inactive or abandoned timelines fade to a muted grey.
- Merge Strategy: Once a branch reaches a conclusion, the AI can summarize the decision and "merge" it back into the main project timeline.
Cmd + Shift + B to create a new branch from the currently selected message.
Auto-Sync Engine
The most powerful feature of Lacesse Intelligence is its ability to write to the database. It doesn't just suggest code; it structures work.
When the AI detects actionable items in the chat (e.g., "We need to update the User Model"), it parses the intent and constructs a JSON representation of a task card.
Ghost Cards
To prevent board clutter, Lacesse Intelligence creates Ghost Cards. These appear on the right-hand board as translucent, 50% opacity items.
- Review: The user sees the ghost card populate in real-time.
- Confirm: Clicking the card solidifies it (100% opacity) and commits it to the database.
- Discard: Ignoring the card causes it to vanish when the chat context changes.
API Reference || Coming Soon
For enterprise integrations, you can interact with the Intelligence engine programmatically.
Analyze Project Context
Sends project metadata to the vector store for indexing.
https://api.lacesse.app/v1/intelligence/index
curl -X POST https://api.lacesse.app/v1/intelligence/index \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"project_id": "prj_89234",
"content": "Refactored the authentication middleware to use JWT.",
"layer": "backend"
}'Generate Task Plan
Forces the AI to generate a structured plan based on a natural language prompt.
https://api.lacesse.app/v1/intelligence/plan
{
// The output format will be compatible with Lacesse Board v2
"goal": "Create a dark mode toggle",
"constraints": ["Must use plain CSS variables", "No flash of unstyled content"],
"response_mode": "json_cards"
}