Neura Market
prompt logo

prompt

Free

Self-wiring, entity-centric knowledge brain for a personal AI agent

FreeFree tier
Inputs: text
Type
Open Source
Company
ai-boost
Visit Website GitHub (40 stars)

About prompt

The Personal Agent Brain Architect is a design blueprint for constructing a self-wiring, entity-centric knowledge brain for a personal AI agent. It ingests meetings, emails, articles, voice notes, and original ideas, then enriches every person and company encountered, fixes citations, and consolidates memory overnight. Unlike generic RAG pipelines or static wikis, it creates a living knowledge graph that extracts typed relationships using deterministic parsers (no LLM calls), ranks by backlink-boosted relevance, and answers compositional questions that vector search alone cannot reach. The design adheres to non-negotiable principles: self-wiring via entity-reference extraction, hybrid search with exact match/graph traversal/vector fallback, verbatim storage with tiered enrichment, scheduled self-maintenance, brain-first lookup before external APIs, and skill-based operation through markdown workflow files.

Key Features

Self-wiring entity-reference extraction using deterministic parsers (no LLM calls)
Typed link creation (attended, works_at, invested_in, founded, advises)
Hybrid search: exact slug match, graph traversal from known entities, vector similarity fallback
Backlink-boosted relevance ranking instead of raw vector similarity
Verbatim ingestion of voice notes and transcripts with no paraphrasing on write path
Tiered enrichment pipeline (Tier 1/2/3) that compiles truth pages without altering originals
Scheduled self-maintenance: citation fixing, stale-page detection, orphan reconciliation, dead-link audits
Dream cycle synthesis: overnight transcript consolidation into reflections and trend identification
Brain-first lookup before external web search/API calls
Skill files (markdown documents) as the interface for operating the brain

Pros & Cons

Pros
  • Self-wiring eliminates the need for LLM calls during entity extraction, reducing cost and latency
  • Hybrid search with backlink boosting outperforms pure vector search (proven benchmark)
  • Verbatim storage preserves original context for voice notes and transcripts
  • Self-maintenance (cron jobs) ensures the brain stays up-to-date and consistent
  • Brain-first approach reduces dependency on external tools and improves response quality
Cons
  • Requires significant setup and ongoing maintenance (cron jobs, tiered pipelines, skill files)
  • Not a plug-and-play solution; deep understanding of knowledge graph design needed
  • Deterministic parsers may miss nuanced relationships that an LLM could catch
  • Orphan detection and reconciliation can create overhead for very large knowledge bases
  • Only as good as the input data: errors or gaps in ingestion propagate through the graph

Best For

Ingesting and enriching meeting transcripts, emails, and voice notes while the user sleepsBuilding a personal knowledge graph for a knowledge worker tracking contacts, companies, and projectsAnswering compositional queries like 'Who works at X?' or 'What did Y invest in?' using graph traversalConsolidating long-term memory across multiple sources with overnight reflection and trend detectionReducing reliance on external APIs by first querying an internal knowledge brain

FAQ

What is the Personal Agent Brain Architect?
It is a design prompt for building a self-wiring, entity-centric knowledge brain for a personal AI agent. It ingests various sources (meetings, emails, articles, voice notes) and creates a living knowledge graph with hybrid search, backlink-boosted relevance, and scheduled self-maintenance.
How does it differ from standard RAG?
It is not a generic RAG pipeline or static wiki. It uses deterministic parsers for entity extraction (no LLM calls), typed links, backlink-boosted relevance, and graph traversal to answer compositional questions that vector search cannot reach.
Is this a ready-to-use tool?
No. It is a design blueprint or prompt that provides guidelines and philosophy for building such a system. It requires implementation based on the described architecture and principles.