Releases·Tool Release
VEKTOR v1.9.8 adds book and document ingestion with stricter reading checks to its local-first AI agent memory
VEKTOR shipped v1.9.8, described as local-first memory for AI agents. The release adds ingestion of personal books and documents and more scrutiny of what an agent reads. The post is a Medium release note.
0 engagement1 sourceWed, Sep 30, 2026, 11:06 PM
AI Overview
- VEKTOR v1.9.8 extends its local-first agent memory to ingest users' books and documents
- The update adds stricter scrutiny of what the AI reads, per the release announcement on Medium
Entities
Related
Products·launched
VEKTOR — memory system for LLM agents with Desk Chat tools and Faraday reasoning
VEKTOR is a memory layer for LLM agents, updated in v1.9.4 with tool use in its Desk Chat interface and a reasoning upgrade to its Faraday component. It aims to give agents persistent, usable memory across conversations. Target users are developers building AI agents.
VEKTOR is a memory layer for LLM agents, updated in v1.9.4 with tool use in its Desk Chat interface and a reasoning upgrade to its Faraday component. It aims to give agents persistent, usable memory across conversations. Target users are developers building AI agents.
0 engagement1 sourcerssMon, Sep 21, 2026, 05:37 AM
Products·building
Local AI memory layer for personal file retrieval
A fully local memory and retrieval engine that indexes personal files (PDFs, notes, screenshots, code) using embeddings and vector search. It solves the problem of finding information in personal files when the user cannot recall the filename or exact wording. The system runs entirely on-device, prioritizing privacy and offline capability.
A fully local memory and retrieval engine that indexes personal files (PDFs, notes, screenshots, code) using embeddings and vector search. It solves the problem of finding information in personal files when the user cannot recall the filename or exact wording. The system runs entirely on-device, prioritizing privacy and offline capability.
2 engagement1 sourceredditWed, Jul 15, 2026, 05:24 PM
Products·launched
Friday: open-source persistent memory layer for AI coding agents
Friday is a self-hosted, open-source memory layer that lets AI coding agents like Cursor, Claude, and Copilot remember across sessions. It stores architecture decisions, schemas, and preferences via the Model Context Protocol. This solves the problem of coding assistants losing context between sessions.
Friday is a self-hosted, open-source memory layer that lets AI coding agents like Cursor, Claude, and Copilot remember across sessions. It stores architecture decisions, schemas, and preferences via the Model Context Protocol. This solves the problem of coding assistants losing context between sessions.
35 engagement1 sourcegithubWed, Sep 16, 2026, 03:53 PM
Releases·Tool Release
Scribe CLI tool released to build AI agent memory from repos and sessions
A developer launched Scribe, a command-line tool that constructs AI agent memory from code repositories and terminal sessions. The tool aims to help agents maintain context across interactions by ingesting local development data.
A developer launched Scribe, a command-line tool that constructs AI agent memory from code repositories and terminal sessions. The tool aims to help agents maintain context across interactions by ingesting local development data.
1 engagement1 sourcehackernewsFri, Jul 17, 2026, 04:44 AM
Products·building
Local memory layer for coding agents to persist context across sessions
A local-first, review-first memory layer that lets coding agents pick up where they left off by storing facts, decisions, and state in a SQLite database with BM25 search. It solves the problem of long-running agent projects losing context or mixing state between projects, targeting developers building agentic coding workflows.
A local-first, review-first memory layer that lets coding agents pick up where they left off by storing facts, decisions, and state in a SQLite database with BM25 search. It solves the problem of long-running agent projects losing context or mixing state between projects, targeting developers building agentic coding workflows.
1 engagement1 sourceredditThu, Jul 16, 2026, 04:14 AM