Atlas - Mémoire Sémantique Persistante

VérifiéSûr

Stockez et récupérez automatiquement la documentation et le code dans une mémoire vectorielle persistante (Qdrant + Voyage). Résout le débordement de contexte en conservant les connaissances entre les sessions.

Spar Skills Guide Bot
DeveloppementIntermédiaire
2002/06/2026
Claude Code
#semantic-memory#vector-database#context-retrieval#persistent-storage

Recommandé pour

Notre avis

Atlas permet d'ingérer et de rechercher du code et de la documentation dans une mémoire sémantique persistante via Qdrant et les embeddings Voyage.

Points forts

  • Stockage persistant du contexte entre sessions, évitant les pertes d'information.
  • Recherche sémantique rapide avec des embeddings de haute qualité.
  • Ingestion récursive de répertoires et support de multiples formats de fichiers.

Limites

  • Nécessite Qdrant en cours d'exécution localement (Docker) et une clé API Voyage.
  • Le temps d'ingestion initial peut être long pour de grands volumes.
  • La mémoire RAM peut être élevée (1.4 Go pour 1M vecteurs).
Quand l'utiliser

Utilisez Atlas lorsque vous devez conserver des informations contextuelles entre sessions ou rechercher des décisions passées dans un projet.

Quand l'éviter

Évitez Atlas pour des tâches ponctuelles sans besoin de mémoire persistante ou si vous ne pouvez pas exécuter Qdrant localement.

Analyse de sécurité

Sûr
Score qualité90/100

The skill describes local document ingestion into a semantic database using standard bash commands (bun, docker, curl). It does not instruct destructive actions, exfiltration, or execution of untrusted code. The only network interaction is localhost health check and the Voyage API, both legitimate.

Aucun point d'attention détecté

Exemples

Remember API architecture
Remember the API architecture we discussed in the last session and store it for future reference.
Ingest research files
Ingest all the .atlas research files in the project directory.
Search authentication patterns
Find all mentions of authentication patterns in the ingested codebase.

name: atlas description: Ingest project documentation and code into persistent semantic memory (Qdrant + Voyage embeddings). Use when user wants to remember context across sessions, ingest docs, or search previous work. Requires Qdrant running locally and VOYAGE_API_KEY set. allowed-tools: Bash(bun:*)

Atlas - Persistent Semantic Memory

Atlas provides automatic context ingestion and retrieval using Voyage embeddings + Qdrant vector database. Solves the context overflow problem by storing knowledge persistently across sessions.

Quick Start

Prerequisites

  1. Qdrant running locally:
docker run -d -p 6333:6333 qdrant/qdrant
  1. VOYAGE_API_KEY set (get from https://voyageai.com):
export VOYAGE_API_KEY="your-key-here"
  1. Verify setup:
curl http://localhost:6333/health

Ingesting Context

Store files in Atlas memory for persistent retrieval:

Ingest Single File

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas ingest /path/to/file.md

Ingest Directory (Recursive)

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas ingest /path/to/docs/ --recursive

Ingest Multiple Paths

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas ingest README.md src/index.ts docs/ -r

What gets ingested:

  • Supported: .md, .ts, .tsx, .js, .jsx, .json, .yaml, .qntm, .rs, .go, .py, .sh, .css, .html
  • Ignored: node_modules, .git, dist, build, coverage, .atlas

Processing:

  • Chunks text (768 tokens, 13% overlap) for semantic coherence
  • Embeds with Voyage-3-large (1024-dim)
  • Stores in Qdrant with dual-indexing (semantic QNTM keys + temporal timestamps)
  • Preserves original text for future consolidation

Searching Context

Retrieve relevant context semantically:

Basic Search

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas search "typescript patterns"

Limited Results

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas search "memory consolidation" --limit 10

Temporal Filtering (Since Date)

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas search "sleep patterns" --since "2025-12-25"

Chronological Timeline

cd ~/production/atlas
bun run --filter @inherent.design/atlas atlas timeline --since "2025-12-01"

When to Use This Skill

Use Atlas when:

  • User asks to "remember this across sessions"
  • Project context is too large for single session
  • User wants to search previous work/decisions
  • Documentation needs to be queryable
  • Building on previous research or code

Examples:

  • "Remember the API architecture we discussed"
  • "What did we decide about the database schema?"
  • "Find all mentions of authentication patterns"
  • "Ingest all the .atlas research files"

Architecture

Built on .atlas research (Steps 1-4 + Sleep Patterns):

Stack:

  • Voyage-3-large embeddings (1024-dim, 9.74% better than OpenAI)
  • Qdrant HNSW index (M=16, int8 quantization, 4x compression)
  • RecursiveCharacterTextSplitter (semantic boundaries)
  • Dual-indexing (QNTM semantic keys + RFC 3339 timestamps)

Production Config (from Step 3 research):

  • Recall@10: >0.98
  • Latency: 10-50ms (p95)
  • Memory: 1.4GB RAM + 5GB disk per 1M vectors

Technical Details

For implementation details, see:

Packages:

  • @inherent.design/atlas-core - Core library (embeddings, storage, search)
  • @inherent.design/atlas - Command-line interface
  • @inherent.design/atlas-mcp - MCP server for Claude Code integration
Skills similaires