Memory Pipeline

Give your AI agent a memory that actually works.

AI agents wake up blank every session. Memory Pipeline fixes that — it extracts what matters from past conversations, connects the dots, and generates a daily briefing so your agent starts each session primed instead of clueless.

What It Does

Why This Is Different

Most "memory" solutions are just vector search over chat logs. This is a cognitive architecture — inspired by how human memory actually works:

• - Extraction over accumulation — Instead of dumping everything into a database, it identifies what's worth remembering: decisions, preferences, learnings, commitments. The rest is noise.
• Knowledge graph, not just embeddings — Facts get linked to each other with bidirectional relationships. Your agent doesn't just find similar text — it understands that a decision about your tech stack relates to a project deadline relates to a preference you stated three weeks ago.
• Briefing over retrieval — Rather than hoping the right context gets retrieved at query time, your agent starts every session with a curated cheat sheet. Active projects, recent decisions, personality reminders. Zero cold-start lag.
• No mid-swing coaching — Borrowed from performance psychology. Corrections happen between sessions, not during. The after-action review feeds into the next briefing. The loop is closed — just not mid-execution.

Quick Start

Install

CODEBLOCK0

Setup

CODEBLOCK1

The setup script will detect your workspace, check dependencies (Python 3 + any LLM API key), create the memory/ directory, and run the full pipeline.

Requirements

• - Python 3
• At least one LLM API key (auto-detected):

Run Manually

CODEBLOCK2

Automate via Heartbeat

Add to your HEARTBEAT.md for daily automatic runs:

CODEBLOCK3

Import External Knowledge

Already have years of conversations in ChatGPT? Import them so your agent knows what you know.

ChatGPT Export

CODEBLOCK4

What it does:

• - Parses ChatGPT's conversation tree format
• Filters out throwaway conversations (configurable: --min-turns, --min-length)
• Supports topic exclusion (edit EXCLUDE_PATTERNS to skip unwanted topics)
• Outputs clean, dated markdown files to INLINECODE12
• Files are automatically indexed by OpenClaw's semantic search

Options:

• - --dry-run — Preview without writing files
• INLINECODE14 — Skip all filtering
• INLINECODE15 — Minimum user messages to keep (default: 2)
• INLINECODE16 — Minimum total characters (default: 200)

Adding Other Sources

The pattern is extensible. Create ingest-<source>.py, parse the format, write markdown to memory/knowledge/<source>/. The indexer handles the rest.

How the Pipeline Works

Stage 1: Extract

Script: INLINECODE19

Reads daily notes (memory/YYYY-MM-DD.md) and session transcripts, then uses an LLM to extract structured facts:

CODEBLOCK5

Output: INLINECODE21

Stage 2: Link

Script: INLINECODE22

Takes extracted facts and builds a knowledge graph:

• - Generates embeddings for semantic similarity
• Creates bidirectional links between related facts
• Detects contradictions and marks superseded facts
• Auto-generates domain tags

Output: memory/knowledge-graph.json + INLINECODE24

Stage 3: Briefing

Script: INLINECODE25

Generates a compact daily briefing (< 2000 chars) combining:

• - Personality traits (from SOUL.md)
• User context (from USER.md)
• Active projects and recent decisions
• Open todos

Output: BRIEFING.md (workspace root)

Performance Hooks (Optional)

Four lifecycle hooks that enforce execution discipline during sessions. Based on a principle from performance psychology: separate preparation from execution.

CODEBLOCK6

Configuration

CODEBLOCK7

Hook Details

Output Files

Customization

• - Change LLM models — Edit model names in each script (supports OpenAI, Anthropic, Gemini)
• Adjust extraction — Modify the extraction prompt in memory-extract.py to focus on different fact types
• Tune link sensitivity — Change the similarity threshold in memory-link.py (default: 0.3)
• Filter ingestion — Edit EXCLUDE_PATTERNS in ingest-chatgpt.py for topic exclusion

Troubleshooting

See Also

• - Setup Guide — Detailed installation and configuration