A single-file HTML dashboard for teams running multiple AI agents in production.
No backend, no signup, no install. Just save index.html and open it in your browser.
If you run more than one AI agent, you have probably asked yourself:
- How much are all these agents costing me this month?
- Which agent is running hot right now?
- What did my agents just do that needs a human review?
- Where is the audit trail when something goes wrong?
Most teams answer these questions with a spreadsheet, a Slack thread, or by logging into three different LLM provider dashboards. AgentOps Mini is a small, opinionated UI for the information you actually need in one place.
Live demo: https://mingyue-ai.github.io/agentops-mini/
Or open index.html in your browser locally. It shows:
- Connected agents and their live status.
- Month-to-date cost across all agents.
- Items pending human review, with one-click approve / reject.
- Per-agent cost history over the last 7 days.
Everything is mocked in static data so you can see the UI immediately.
We wanted to validate whether the visual part of an agent-operations dashboard is useful before building auth, integrations, or a backend. If this view resonates with you, the next step is real data from your agent runs.
What is the hardest part of running multiple agents in production?
- Tracking costs across providers?
- Reviewing outputs before they go live?
- Knowing which agent is doing what?
- Something else entirely?
Tell us in an issue or in the comments wherever you found this.
If you are interested in a hosted or self-hosted version with real data, say so in an issue. We are talking to teams about what an agent-operations layer should actually include.
MIT