Hindsight Documentation FAQs

Question 1

What are Dispositions in Hindsight?

Accepted Answer

Dispositions are configurable traits like skepticism, literalism, or empathy (rated 1-5) that influence how the AI agent reflects on and reasons with retrieved memories.

Question 2

Which retrieval strategies does Hindsight use?

Accepted Answer

Hindsight utilizes four parallel strategies: semantic (vector) search, BM25 (keyword) search, graph-based retrieval for relationships, and temporal search for time-based context.

Question 3

What is Hindsight?

Accepted Answer

Hindsight is a biomimetic memory system for AI agents that organizes information into memory banks and uses four parallel retrieval strategies to help agents learn and recall facts over time.

Question 4

How can I integrate Hindsight into my project?

Accepted Answer

You can integrate Hindsight using its specialized SDKs for Python, Node.js, or Rust, or interact directly with its REST API after deploying the server via Docker or Kubernetes.

Question 5

How does Hindsight manage conversation history?

Accepted Answer

Hindsight uses a 'document_id' to group messages in a conversation, allowing for efficient upserts and evolution of memories as the conversation progresses.

Hindsight Documentation

Key Features

Use Cases

Hindsight Documentation

Key Features

Use Cases