If you only read one page, read this.

Install

git clone https://github.com/Hulupeep/pastchats-memory.git
cd pastchats-memory
python3 -m venv .venv
source .venv/bin/activate
pip install -e .

Create the memory DB

pastchats-memory init --db .swarm/prompt_memory.db

Index your history

Point this at whatever folder(s) contain your chat exports or coding-agent logs.

pastchats-memory index --db .swarm/prompt_memory.db --input ~/projects

Index multiple places:

pastchats-memory index --db .swarm/prompt_memory.db --input ~/projects ~/Downloads/chat_exports

Only index user+assistant:

pastchats-memory index --db .swarm/prompt_memory.db --input ~/projects --roles user,assistant

Recall before you code

Write your query like:

<problem> + <constraint> + <stack>

Example:

pastchats-memory recall --db .swarm/prompt_memory.db --query "webhook retry idempotent python"

If recall is empty

Do this:

• Make the query more specific (include stack + constraint)
• Index more history
• Proceed normally (don’t force it)

Optional speed boost (sqlite-vec)

If you installed sqlite-vec:

export PROMPT_MEMORY_SQLITE_VEC_PATH=/absolute/path/to/vec0.so

Then re-run index.

Optional better semantic search (OpenAI embeddings)

pip install -e .[openai]
export OPENAI_API_KEY=your_key
pastchats-memory index --db .swarm/prompt_memory.db --input ~/projects --embed-provider openai