by JasonDocton
Memory for AI that works like yours—local, instant, persistent. 13x faster than Pinecone, 5x leaner than RAG. Finds what RAG misses. Zero cloud, zero cost.
# Add to your Claude Code skills
git clone https://github.com/JasonDocton/lucid-memory2.7ms retrieval. 743,000 memories/second. $0/query.
Memory for AI that works like yours—local, instant, persistent.
curl -fsSL https://lucidmemory.dev/install | bash
New in 0.6.0: <a href="#memory-consolidation">Memory Consolidation</a> — Lucid Memory is now self-maintaining. Background consolidation strengthens recent memories, decays stale ones, prunes weak associations, and manages visual memory lifecycle. New memories are checked against existing traces — similar content reinforces or updates rather than duplicating. 307 tests, 0 tsc errors.
New in 0.5.0: <a href="#episodic-memory">Episodic Memory</a> — Claude remembers not just what happened, but how it unfolded — reconstructing the story of your debugging session, not just the fix. "What was I working on before the auth refactor?" now has an answer.
We're not a vector database. We're the retrieval layer that makes vector databases obsolete for AI memory.
Pinecone stores vectors. We understand context.
<table> <tr> <th></th> <th>Lucid Memory</th> <th>Claude-mem</th> <th>Pinecone RAG</th> <th>Traditional RAG</th> </tr> <tr> <td><b>Retrieval Speed</b></td> <td>✅ <b>2.7ms</b></td> <td>~50ms</td> <td>10-50ms</td> <td>200-800ms</td> </tr> <tr> <td><b>Token Efficiency</b></td> <td>✅ <b>5x</b></td> <td>1x (baseline)</td> <td>2.5x</td> <td>~2x</td> </tr> <tr> <td><b>Recall @ Fixed Budget</b></td> <td>✅ <b>82.5%</b></td> <td>28.9%</td> <td>55.3%</td> <td>~50%</td> </tr> <tr> <td><b>Storage Compression</b></td> <td>✅ <b>5x (80% smaller)</b></td> <td>1x</td> <td>1x</td> <td>1x</td> </tr> <tr> <td><b>Query Cost</b></td> <td>✅ <b>$0</b></td> <td>$0</td> <td>$70+/month</td> <td>API costs</td> </tr> <tr> <td><b>Recency vs Relevance</b></td> <td>✅ <b>Multiplicative (relevance wins)</b></td> <td>Binary 90-day filter</td> <td>No recency</td> <td>No recency</td> </tr> <tr> <td><b>Associative Retrieval</b></td> <td>✅ <b>3-hop spreading activation</b></td> <td>None</td> <td>None</td> <td>None</td> </tr> </table>No comments yet. Be the first to share your thoughts!
<sub>Benchmarked on realistic developer workflows (50-200 memories). Full methodology: <code>bun run bench:realistic && bun run bench:tokens</code></sub>
| | Lucid Memory | Pinecone | |---|---|---| | Token efficiency | 5x | 2.5x | | Recall | 82.5% | 55.3% | | Latency | 2.7ms | 10-50ms | | Monthly cost | $0 | $70+ | | Your data | Stays on your machine | Sent to cloud | | Recency awareness | Yes (multiplicative) | No | | Associative retrieval | Yes (spreading activation) | No |
Pinecone is a great vector database. But vector search isn't memory.
Lucid Memory retrieves 50% more relevant context (82.5% vs 55.3% recall), runs 10-20x faster (local vs cloud), costs nothing, and keeps your code *...