Standalone hybrid-search memory backend for Cortex. Postgres + pgvector + tsvector, fused with reciprocal rank fusion. Cortex's canonical storage backend — supports embedded PGlite for personal installs and external Postgres for VPS / cloud deploys. Self-
Standalone hybrid-search memory backend for Cortex. Postgres + pgvector + tsvector, fused with reciprocal rank fusion. Cortex's canonical storage backend — supports embedded PGlite for personal installs and external Postgres for VPS / cloud deploys. Sel
Doc pipeline: chunk prose content by heading, produce ingestible memories.
Code pipeline: per-file memories with language detection and heuristic chunking.
Conversation pipeline: chat threads and email chains into searchable memories.
3-pass meeting transcript extraction pipeline (structural -> synthesis -> brief).
przm Cortex MCP server: unified work-knowledge layer.
pgvector support for Node.js, Deno, and Bun (and TypeScript)
Shared types and interfaces for Cortex adapters and pipelines.
Bitbucket Cloud source adapter for Cortex.
Slack channels + threads source adapter for Cortex.
Linear source adapter for Cortex (GraphQL).
Base classes and helpers for building Cortex source adapters.
GitHub source adapter for Cortex.
Atlassian Confluence source adapter for Cortex.
OpenRouter LLM provider for Cortex (BYOK cloud aggregator).
Local Ollama LLM provider for Cortex.
Atlassian Jira source adapter for Cortex.
Generic multi-stage pipeline framework for Cortex.
Obsidian vault source adapter for Cortex.
Notion source adapter for Cortex.
Base classes and helpers for building LLM provider packages.
Loom recordings + transcripts source adapter for Cortex.
LLM provider interface and per-task router for Cortex.