Agentic Data Stack#171
Conversation
There was a problem hiding this comment.
You run it yourself
You can, but to reduce friction we should lead with the fully managed service on cloud while still highlighting that it's simple for users to self host
There was a problem hiding this comment.
Updated the doc to now lead with ClickHouse Agents and link out to the docs, following with self-hosted Agentic Data Stack. Let me know if you want to tweak further or shift messaging entirely.
There was a problem hiding this comment.
needs feedback loop ability, via langfuse
…into agentic-data-stack
|
@dhtclk also compare with https://langfuse.com/integrations/agentic-data-stack |
|
Thanks @graphaelli - made some tweaks based on feedback. Compared to and linked out to langfuse docs as well. |
…into agentic-data-stack
| ## What you can do {#what-you-can-do} | ||
|
|
||
| - **Ask questions in natural language** and get answers drawn from your own data. | ||
| - **Build agents with no code:** give an agent instructions and tools, then reuse it. |
There was a problem hiding this comment.
| - **Build agents with no code:** give an agent instructions and tools, then reuse it. | |
| - **Build agents with no code** by giving an agent instructions and tools, then reusing it. |
| - **Build agents with no code:** give an agent instructions and tools, then reuse it. | ||
| - **Share agents and conversations** as read-only links, so others can trace the queries behind an answer. | ||
| - **Generate interactive charts and visualizations** from query results inside a conversation. | ||
| - **Evaluate and improve answers:** score responses in Langfuse with human review or an LLM judge, then refine your prompts and agents. |
There was a problem hiding this comment.
| - **Evaluate and improve answers:** score responses in Langfuse with human review or an LLM judge, then refine your prompts and agents. | |
| - **Evaluate and improve answers** by scoring responses in Langfuse with human review or an LLM judge, and refining your prompts and agents. |
|
|
||
| The agent works through the ClickHouse MCP server to inspect your databases and tables, run read-only queries, and build an answer from the results. The stack wires this up for you, so LibreChat queries your data from the first sign-in. Stand up the full stack with the [Docker setup guide](/products/agentic-data-stack/docker-setup). | ||
|
|
||
| ## Build an agent over your data {#build-an-agent} |
There was a problem hiding this comment.
This page feels a little weird with an H2 header for every capability of LibreChat. Maybe it would be better suited as a table under an H2 header called "Librechat capabilities"/"Librechat features" or something similar.
There was a problem hiding this comment.
Maybe it's fine, not sure, will leave up to you.
3 participants
Iterating on Agentic Data Stack Docs
Note
Low Risk
Documentation-only changes with no application or infrastructure code.
Overview
Replaces the Agentic Data Stack placeholder with a full documentation set: a landing overview (managed Cloud vs self-hosted, architecture diagram, capabilities, component table), a Docker Compose walkthrough (
prepare-demo.sh, ports, MCP selection, teardown/reset, compose wiring), and three component guides for LibreChat, the ClickHouse MCP server, and Langfuse.Navigation is reorganized from a single Overview group into Get started (overview + docker-setup) and Components (the three component pages).
Reviewed by Cursor Bugbot for commit b6a8785. Bugbot is set up for automated code reviews on this repo. Configure here.