Skillget

aletheia-platform

mcp
Guvenlik Denetimi
Uyari
Health Uyari
  • No license — Repository has no license file
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Uyari
  • process.env — Environment variable access in inspect_keys.js
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

AletheiaDB: The cognitive memory engine for AI agents with fact supersession.

README.md

AletheiaDB Platform

AletheiaDBDB is a temporal memory database for AI agents. It stores memories as evolving evidence, tracks which facts are currently true, invalidates stale facts, computes numeric answers deterministically, and retrieves context through a hybrid of vector, lexical, graph, and temporal search.

Unlike generic memory APIs that just wrap embeddings and return stale information, AletheiaDB is an actual database engine built in Rust. It focuses on the core problem of long-term agent memory: knowing what is currently true, rather than just recalling what was said in the past.

This open-source repository (aletheia-platform) powers the official web platform deployed at aletheiadb.com, built with Qwik. It serves the marketing landing page, developer documentation, blog, and the user dashboard for API key management.

What AletheiaDB Does Better (The Benefits)

  • Temporal Truth & Fact Supersession: AletheiaDB understands when a new fact supersedes an old one (e.g., "I moved to Seattle" invalidates "I live in Austin"). Stale facts are filtered out, giving your agents accurate context.
  • Deterministic Numeric Memory: It computes numeric answers (counts, sums) deterministically using a metric vault, rather than relying on the LLM to guess the right number from a context window.
  • Local-First, Single-Binary Engine: The core engine is deployed via a highly performant Rust binary. Keep your data private and local.
  • True Hybrid Retrieval: Combines HNSW vector search, BM25 full-text search, graph knowledge retrieval, and time-aware ranking in one unified system.
  • Drop-in Proxy: Add memory to existing OpenAI SDK apps by simply changing the base URL. AletheiaDB automatically injects context and forwards the request to your LLM provider.
  • Model Context Protocol (MCP): Designed for IDEs and agents like Claude Code, Cursor, and Windsurf via an MCP stdio server.

How to Use AletheiaDB

1. Local Engine

Run AletheiaDB privately on your local machine for privacy-focused coding agents.

aletheia serve

2. Drop-In Proxy

For developers already using OpenAI-style APIs, add memory to your agents without rewriting your application code. Change your SDK's base URL to point to your AletheiaDB instance, and pass your AletheiaDB API key.

3. Hosted Platform & API Keys

The hosted API provides managed memory infrastructure with API keys and usage stats. Log into the platform dashboard to manage your credentials.

Developing the Web Platform

This repository contains the Qwik rewrite of the AletheiaDB dashboard and marketing site.

Includes

  • Marketing landing page on /
  • Documentation surface on /docs
  • Blog surface on /blog
  • Demo login on /login
  • Cookie-backed API key dashboard on /platform

Run locally

npm install
npm run dev

Production build

npm run build
npm run serve

Demo Login

You can test the platform dashboard locally using the demo credentials:

(Override these by setting ALETHEIADB_DEMO_EMAIL and ALETHEIADB_DEMO_PASSWORD in your environment.)

Blog Authoring

Blog posts live in src/content/blog/*.md.

  • The filename becomes the slug.
  • Frontmatter drives title, description, dates, tags, and featured state.
  • New .md files are picked up automatically by the blog index, post routes, sitemap, and SEO metadata.

Yorumlar (0)

Sonuc bulunamadi