besslframework-stack/project-tessera

Overview

Tessera is a local-first memory layer for AI agents, providing 58 MCP tools and 54 REST endpoints for persistent context storage. All data is encrypted at rest with AES-256, ensuring privacy without relying on external cloud services.

Best for

Best for
Developers needing a simple, local, encrypted memory layer for AI agents and chatbots

Use cases

• Adding persistent memory to AI agents for continuity across sessions
• Storing and retrieving encrypted context via MCP or REST APIs
• Building privacy-preserving applications that keep data local

How to use

Install

uvx --from project-tessera tessera setup

Tools exposed

• search_documents
• unified_search
• view_file_full
• read_file
• list_sources
• list_memories
• forget_memory
• export_memories
• import_memories
• memory_tags
• search_by_tag
• memory_categories
• search_by_category
• find_similar
• knowledge_graph
• digest_conversation
• toggle_auto_learn
• review_learned
• session_interactions
• recent_sessions

Tested with

Claude Desktop, ChatGPT

Notes

Tessera is a local-first memory layer for AI agents, providing 58 MCP tools and 54 REST endpoints for persistent context storage. All data is encrypted at rest with AES-256, ensuring privacy without relying on external cloud services.

14 stars on GitHub. Last updated 2026-03-21. Licensed AGPL-3.0.

Use cases

• Adding persistent memory to AI agents for continuity across sessions
• Storing and retrieving encrypted context via MCP or REST APIs
• Building privacy-preserving applications that keep data local

Pros

• Local-first architecture eliminates dependency on third-party cloud memory
• Offers both MCP tools and REST endpoints for flexible integration
• Encrypts all stored data with AES-256 for security

Cons

• Small community and low GitHub stars indicate early-stage project
• Limited documentation and third-party tooling
• Potential performance bottlenecks for high-throughput use cases due to local encryption

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.