explaingit

1lystore/awaek

13Python

TLDR

Awaek is a small skill for a chat assistant called Hermes Agent that lets you talk to your own saved X (formerly Twitter) bookmarks.

Mindmap

A visual breakdown will appear here once this repo is fully enriched.

In plain English

Awaek is a small skill for a chat assistant called Hermes Agent that lets you talk to your own saved X (formerly Twitter) bookmarks. The pitch in the README is that most people save good posts and then never use them again, so Awaek pulls those saves into a local searchable library that the chat agent can answer questions from, draft posts with, or use to put together plans. The sample interactions show the kind of thing you can ask. Examples include asking what your saved posts say about AI agents, finding a specific old bookmark about x402 payments, turning launch-related bookmarks into a 30-day plan, or drafting twenty new posts in the style of your saved growth posts. The README is direct that everything stays on your machine: there is no hosted backend, no cloud database, no pasted X secrets, and authentication to X is delegated to a separate tool called xurl. Installation is one line. Inside any Hermes chat (terminal, Telegram, Slack, Discord, or the Open Web client) you can say install Awaek from 1lystore/awaek/skills/awaek, or run hermes skills install with the same path from a terminal. Awaek needs Hermes Agent, Python 3, and xurl authenticated against an account that can read its own bookmarks. The README links to the official X Developers post that explains setting up xurl, and gives a one-line curl install for it. Once installed, Awaek sync pulls bookmarks down (up to 100 recent saved posts in the current version), stores them in a local SQLite database at ~/.hermes/awaek/data/awaek.db, categorizes them, and builds a local search index. After that you can ask things like Awaek, show my bookmark topics or Awaek, find my saved posts about pricing. Under the hood the README says it syncs via xurl, stores in SQLite, categorizes, and builds evidence packs that the Hermes Agent can cite when answering. The roadmap mentions improvements like smarter topic learning, better packing of evidence for large bookmark libraries, a sample mode that works without a real X setup, connectors to saved-content sources beyond X, and richer writing workflows. The repository topics list it under ai-agent and hermes-skill tags.

Open on GitHub → Explain another repo

Generated 2026-05-21 · Model: sonnet-4-6 · Verify against the repo before relying on details.