Hey HN! We’re building Athena, an AI agent that does data analysis, text-to-SQL, multi-modal whiteboard-style data transforms, python execution environment, and more. I wanted to share a capability from our latest platform release that we’re excited about: diagrams!
Mermaid diagrams have been making their rounds recently, so we gave them a try and it turned out… pretty awesome. You can ask Athena to generate ERDs (entity relationship diagrams) when documenting datasets, flowcharts when explaining analyses, queries, data quality issues, code execution, etc. It works in a much wider range of use cases though: we asked Athena to generate a visual primer on Marvel lore, a web service schema, customer journey map for PepsiCo and got impressive results. We also added a live editor, so even if the generation fails, you can inspect the Mermaid code and fix it live.
We’re very impressed with the way an AI agent can work with diagrams, because it feels like a step-up from text generation (even if it’s still just text). It focuses on concepts, relationships between entities, and ontologies. And, as someone mentioned on HN, it allows bi-directional editing (both users and agent can edit input and output). We’re bullish about enabling enterprise data teams to move from semantic model planning to execution in minutes.
Give it a try yourself — Athena has a free tier (no credit card sign-up). You can give Athena a link to your website, github repo, product and ask her to visualize it in a mermaid diagram or upload a dataset and create visual documentation. Would love to hear your feedback and see notable diagram outputs in the thread!
Also, once you have an account try forwarding an email to her ([email protected]), then navigate to conversations to see the results.
How to get to diagram generation:
register → go to chat → start a chat
or find a “generate diagrams” task on a homepage
Mermaid diagrams have been making their rounds recently, so we gave them a try and it turned out… pretty awesome. You can ask Athena to generate ERDs (entity relationship diagrams) when documenting datasets, flowcharts when explaining analyses, queries, data quality issues, code execution, etc. It works in a much wider range of use cases though: we asked Athena to generate a visual primer on Marvel lore, a web service schema, customer journey map for PepsiCo and got impressive results. We also added a live editor, so even if the generation fails, you can inspect the Mermaid code and fix it live.
You can check out a quick demo here: https://youtu.be/BSB0IoUCE0s
We’re very impressed with the way an AI agent can work with diagrams, because it feels like a step-up from text generation (even if it’s still just text). It focuses on concepts, relationships between entities, and ontologies. And, as someone mentioned on HN, it allows bi-directional editing (both users and agent can edit input and output). We’re bullish about enabling enterprise data teams to move from semantic model planning to execution in minutes.
Give it a try yourself — Athena has a free tier (no credit card sign-up). You can give Athena a link to your website, github repo, product and ask her to visualize it in a mermaid diagram or upload a dataset and create visual documentation. Would love to hear your feedback and see notable diagram outputs in the thread!
Also, once you have an account try forwarding an email to her ([email protected]), then navigate to conversations to see the results.
How to get to diagram generation: register → go to chat → start a chat or find a “generate diagrams” task on a homepage