An always-on ambient intelligence layer that builds a continuous knowledge graph of your conversations and projects without needing explicit commands.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install percept-ambient
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install percept-ambient using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is percept-ambient?
percept-ambient is a high-performance background service designed to give AI agents persistent situational awareness. By passively processing ambient information, it constructs a sophisticated knowledge graph of entities, relationships, and project statuses. This skill eliminates the need for manual context setting, allowing your agent to remain synchronized with your real-world activities through the power of Openclaw Skills.
Built for privacy and speed, it summarizes ongoing interactions to ensure that when you ask your agent a question, it already has the necessary background data. It functions as a bridge between raw audio input and actionable agent intelligence, making it an essential component for any advanced autonomous setup.
percept-ambient Use Cases
- Maintaining always-on context awareness for seamless agent-human interaction.
- Building an automated background knowledge base from daily professional conversations.
- Querying agents about specific people, projects, or decisions based on overheard context.
- Reducing token usage by providing dense context packets instead of full chat histories.
How percept-ambient Works
- The system continuously captures conversations and passes them through a summarization pipeline.
- Key entities such as people, companies, projects, and topics are automatically extracted.
- Semantic relationships are mapped between entities (e.g., who is working on which client project).
- Rich context packets are assembled on demand whenever an agent action is triggered.
- Retrieval is handled through a hybrid approach using SQLite FTS5 and LanceDB vector search.
percept-ambient Setup
To utilize this skill, you must first have the percept-listen and percept-summarize skills active. Ensure your environment is configured for local data storage.
# Monitor the service or perform a hybrid search test
curl localhost:8960/api/search?q=project+deadline&mode=hybrid
Access the real-time management dashboard at http://localhost:8960 to configure privacy settings and view the entity graph.
percept-ambient Data Schema & Taxonomy
The skill organizes data into highly structured Context Packets to optimize agent performance. Storage is managed locally to ensure privacy within the Openclaw Skills ecosystem.
| Component | Description |
|---|---|
| recent_conversations | A curated list of the latest interaction summaries |
| resolved_entities | Unique identifiers for people, projects, and organizations |
| relationships | Directed links showing how entities connect (e.g., works_on) |
| relevant_history | Historically significant data points retrieved via vector search |
percept-ambient Advanced Features
- Hybrid search architecture combining FTS5 text search and LanceDB vector retrieval.
- Native support for NVIDIA NIM embeddings with local all-MiniLM-L6-v2 fallback.
- Granular privacy controls including configurable TTL auto-purge and transcript-only storage.
- Real-time entity graph visualization for monitoring agent knowledge state.
- Seamless integration with other Openclaw Skills for automated knowledge management.
SKILL.md
Loading
Related Openclaw Skills
Browser Audio Capture
A bridge that enables AI agents to capture and process real-time audio streams from any browser tab without platform-specific integrations.
watchclaw
An automated health monitoring and recovery watchdog designed to protect OpenClaw gateways from faulty configuration changes.
ShieldCortex
ShieldCortex provides AI agents with a secure, persistent memory system featuring semantic search, knowledge graphs, and a 6-layer defense pipeline.
Seats.aero Award Flight Search
Search and find award flight availability across 24 mileage programs using the seats.aero partner API.
Percept Listen
Ambient audio capture and local transcription for AI agents via wearable devices like Omi and Apple Watch.
Percept Speaker ID
A sophisticated speaker identification system that maps anonymous transcript labels to real names and manages voice command authorization levels.
Featured*
