building @runtypelabs • 🤠→🌉 • 🧠 APIs, evals, carcinization
San Francisco
- Tweets 152
- Following 65
- Followers 47
- Likes 67
Your code mode composition makes sense from your position as a gateway, but I think there are other contexts where it’s almost essential when you have a large number of tools. Especially if you’re trying to get into plugin marketplaces that punish large numbers of tools (like >20)
11
I just figured they started running quants after the initial PR win from the new model + when they need to claw the GPUs back for training the next one
261
Hilariously overbuilt `caffeinate` shortcut, which responds to ambient conditions (system load = coffee temp) and your macbook's motion sensors
github.com/bookernath/caf
2
177
Big fan of @mattpocockuk 'grill me' skill but the new version asks a bunch of questions at once, which is better but also means a lot of typing and it doesn't fit inside Claude Code's AskUserQuestion limits. So I built a UI for it that uses MCP Elicitation to pop up.
github.com/bookernath/grill-…
1
1
5
103
Cloud platforms gave developers reliable primitives for compute, unlocking tremendous productivity and giving birth to the modern SaaS era. Agentic products need a different set of tools in order to get maximum leverage out of LLMs; models by themselves are unrefined ore, waiting to be formed into useful things by the right forge.
I wrote about how we think about this emerging layer at Runtype, and propose a set of the new primitives:
runtype.com/blog/new-primiti…
1
3
68
Ahead of our @Cloudflare hackathon tomorrow: how Runtype makes agents on other frameworks (here, @flueai) better over time.
Building on Flue and want a faster improvement loop? runtype.com
1
1
2
79
Cold startup time on Cloudflare workers IS harmed by a larger bundle, even if it has dead or lazy code. This is not expressed by the startup time metric in wrangler.
1
2
65
I built an AI data visualization agent that turns a plain-English question into an interactive chart, data table, and auditable SQL. Works on Web, Slack, Email, or wherever you want.
Uses a @RuntypeLabs agent, @insforge for DB + app hosting, and Flint from @Microsoft for Gen UI
4
2
13
99,965
Persona is live on Product Hunt today - if you've gotten value from it, give it an upvote!
producthunt.com/products/per…
4
5
72
You can build agentic experiences *on* your site using WebMCP today with github.com/runtypelabs/perso… which includes the polyfill for other browsers and sites not in the origin trial
1
2
176
Runtype (my platform), which is mostly using cloudflare's infra primitives under the hood. Big difference being we've made agent building much less codeful than using their stuff directly. Much easier for non-engineers.
2
489
These are such great points. I was happy to see Persona already supported all of them, but this encouraged me to adjust our default scroll behavior to to `anchor-top`
persona-chat.dev/scroll-engi…
What Makes a Great Streaming Chat Experience
Start here: Never move the reader against their intent.
1. Move only when the reader asked to move.
If someone is reading, don’t pull them somewhere else. Auto-scroll should never be the default.
2. Follow only while they’re following.
If they’re at the live edge, keep the stream in view. If they scroll away, leave them there.
3. Every interaction is intent.
Scrolling is not the only signal. Selecting text, using the keyboard, opening a link, searching should all stop the interface from moving.
4. Start a new turn near the top of the viewport.
This gives the new turn somewhere it can be read from the beginning.
5. Then stream in the answer.
The streaming answer can then grow into the available space.
6. Keep part of the previous conversation in context.
Enough of the previous turn should remain visible so the reader knows where they are. Context.
7. Let new content arrive offscreen.
The conversation can keep streaming without changing what the reader is looking at.
8. Show what’s happening out of view.
Make it clear when a response is still streaming or when new messages have arrived.
9. Make it easy to return to the latest reply.
A "Jump to latest" action should bring the reader back and resume following.
10. Let people jump anywhere in the conversation.
Long threads need message links, search, unread markers, and direct navigation.
11. Reopen where the reader left off.
A saved conversation should open at the last meaningful turn. This is usually the last user message. Not the absolute bottom.
12. Keep the reader’s place when layout changes.
Images load. Markdown expands. Code blocks render. Older messages appear above. None of that should make the reader lose their place.
13. Handle interruptions without stealing position.
Stopping, retrying, regenerating, branching, or errors should not unexpectedly move the conversation.
14. Stay responsive in long threads.
Streaming text, markdown, code, images, and long history should still feel responsive.
15. Be accessible without the noise.
Keep the transcript navigable, preserve keyboard focus, and announce important events at a comfortable pace.
It’s all about the scroll. Scroll Engineering.
3
107