Credentials. The server holds your LLM key in its environment; do not expose it publicly without auth.
Who it is forA developer whose assistant forgets the user between sessions.
The problemHand-rolling memory means deciding what to store, when to search, and how to inject it.
The outcomeThe same chat code, now recalling relevant memories per user bank and saving new ones automatically.
What you need first
- Docker, pip or Kubernetes
- An LLM API key (OpenAI in the quick start)
- Python, Node.js or Go for the client
Steps
- Start a Hindsight server, for example with the Docker command from the README on ports 8888 and 9999
- Install the wrapper:
pip install hindsight-litellm - Import it with
from hindsight_litellm import wrap_openaiand wrap your OpenAI client, passing a bank_id per user and hindsight_api_url for a self-hosted server (http://localhost:8888) - Call
client.chat.completions.createas before; usewrap_anthropic()for the Anthropic SDK
Prompt for your agent
Add vectorize-io/hindsight memory to my chatbot in <file> using the LLM wrapper against a local server, one bank per user id, and show me a two-session test where the second session remembers a fact from the first.
Caveat.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for vectorize-io/hindsight and "Add long-term memory to an existing chatbot in two lines"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer