Agent Docs - Iterate#6288
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
4 Skipped Deployments
|
dhtclk
changed the title
| - **Max Context Tokens** — caps total context the agent sends to the model. Leave as **System** to use the model's default. Lower it to reduce cost; raise it for agents that need to reason over large inputs. | ||
| - **Max Output Tokens** — caps the size of the agent's response. **System** uses the model's default. Set lower if responses are too long, higher if they're being cut off. | ||
| - **File Token Limit** — caps how many tokens a single uploaded file contributes to context. Useful when users attach large files and you don't want them to crowd out the rest of the conversation. | ||
| - **Max Context Tokens** - caps total context the agent sends to the model. Leave as **System** to use the model's default. Lower it to reduce cost; raise it for agents that need to reason over large inputs. |
There was a problem hiding this comment.
I don't see this when I try the model builder, is this just not available in the Open House demo?
There was a problem hiding this comment.
Ah, this is only available for certain models, none of which are enabled in production yet. I'm ok to remove this and add back when those are available.
There was a problem hiding this comment.
I've added draft=true to the frontmatter of this doc so that it doesn't show in production. When these features are available in production we'll just need to remove that attribute and add it to the overview page again.
| - **Temperature** — randomness. Higher values (0.7–1.0) = more random, while lower values (0.0–0.3) = more focused and deterministic. We recommend altering this or Top P but not both. | ||
| - **Top P** — nucleus sampling. Changes how the model selects tokens for output. | ||
| - **Top K** — restricts sampling to the top K most likely tokens at each step. Supported by some providers; controls determinism along a different axis than temperature. | ||
| - **Temperature** - randomness. Higher values (0.7–1.0) = more random, while lower values (0.0–0.3) = more focused and deterministic. We recommend altering this or Top P but not both. |
Blargian
left a comment
Blargian
left a comment
There was a problem hiding this comment.
LGTM, let's wait for any additional feedback from @graphaelli
|
Docs PR opened: ClickHouse/mintlify-docs-dev#73 Ported ClickHouse Agents documentation (16 pages plus screenshots) from the upstream clickhouse-docs repo into the Mintlify AI/ML section. |
…show up right now
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes using default effort and found 3 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit fefa2d2. Configure here.
3 participants
Summary
Follow up PR for #6281
Iterating on docs to get them up to par on quality after the initial pass. Primarily adding screenshots for this pass to start.
Resolves DOC-786
Checklist
Note
Low Risk
Documentation-only changes with no runtime impact; verify screenshot assets exist and fix the broken Memory intro if present.
Overview
This PR iterates on ClickHouse Agents cloud docs (follow-up to the initial Agent Builder pass): it aligns prose with the current console UI, adds screenshots (
IdealImage) for key flows, and tightens section structure.Structure and navigation: The Agent Builder overview is now a landing page (
doc_type: landing-page) with a page table and a quickstart tip, instead of long inline sections on identity and configuration. The agents overview and in this section lists use tables for skimmability.Quickstart: Rebuilt as a
VerticalStepperflow—launch from the service sidebar, create agent fields (including Category), then split Capabilities (Run Code, Web Search, MCP, Skills, etc.) vs Add Tools (third-party/image integrations), with step images through a sample query.Product accuracy: Docs now distinguish Capabilities vs Add Tools—e.g. Run Code / Web Search toggles vs image generation via the tools modal; vision via composer Upload to Provider; sharing via the builder share icon/modal; memory and prompts tied to left-nav panels. Model parameters gains console open steps and is marked
draft: true.UX polish: Chat, marketplace, memory, and prompts pages add step-by-step UI guidance and screenshots (fork modes, multi-conversation, share link, marketplace actions, memory CRUD, prompt preview/use). Minor list punctuation normalized (em dashes to hyphens in several builder pages).
Note for reviewers:
memory.mdopens with a sentence that appears to drop the word Memory (“is a per-user store…”)—likely an accidental edit worth fixing before merge.Reviewed by Cursor Bugbot for commit bec64dc. Bugbot is set up for automated code reviews on this repo. Configure here.