Live web research agent with citations

Glim MCP is an aeon skill: Live-data research via the glim.sh MCP - web search, full page extraction, X/Twitter, Reddit, GitHub, Amazon, and YouTube transcripts - synthesized into a cited digest. Pay-per-call from the connected account balance; OAuth Connect via the dashboard MCP panel. Like every aeon skill it is one SKILL.md prompt file that runs unattended on your own GitHub Actions, on a schedule or when you trigger it, and sends what it finds to Telegram, Discord, Slack or email.

Answer one research question with live data through the glim.sh MCP server (glim.sh/mcp): web search, full-page extraction, and platform-native access to X/Twitter, Reddit, GitHub, Amazon, and YouTube transcripts. Every call draws from the operator's prepaid glim balance - spend is real, so the sweep is bounded.

Every aeon skill takes one optional text input, var, set in aeon.yml or typed in when you run it by hand. It picks the target or the mode for that run.

The research question or task, e.g. what are people saying about MCP servers this week or pull the top HN + Reddit takes on <topic>. Append --deep for a wider sweep. Required. If empty, log GLIM_NO_QUERY and exit cleanly (no notify).

Besides the model you connected to aeon, Glim MCP uses the following. Add secrets as GitHub Actions secrets on your aeon repo, or in Aeon Connect.

• glim MCP server - required, connect it once from the dashboard MCP panel (OAuth)

Easiest: open Aeon Connect, sign in with GitHub and turn on Glim MCP. No clone, no terminal. To add it to an existing aeon repo from a shell, run:

Then switch it on in aeon.yml (the default schedule is filled in):

skills:
  glim-mcp: { enabled: true, schedule: "workflow_dispatch", var: "" }

New to aeon? Read the docs for how skills, schedules and memory fit together, or browse all skills.