WIP: overhaul, prep comps, prep agent kit

[andrewxhill]

@dtbuchholz here is a pretty WIP overhaul. I think it's super promising though, it gives a really nice flow into the information and it feels like the right things are highlighted at the right time. It feels like there are three tasks here:

• review accuracy of new stuff (agent toolkit, mcp). i tried to use some of the agent toolkit included examples, but hit snags so prob better for your eyes.
• simplify and cut stuff that is TMI, not ready, no longer right.
• open tix on enhancements we could make (tests, building off other package docs, etc)

Could you take a spin on this? I left some comments for mark below too, so maybe we can tap him in afterwards?

Overview

A WIP overhaul of docs. Brings them up to speed with competition role and framing. Makes beginner uses the primary focus, since every user is a beginner on recall right now.

• Quickstart and Framework guides that focus on 0-something very quickly
• MCP at the forefront along with agent toolkit.
• Move deeper tech down below
• Adds places for competition info to go public already, including trading sim.

Notes below.

Agent toolkit

@dtbuchholz to review and simplify any here:

• docs/agent-toolkit/*

MCP

@dtbuchholz to review and simplify any here:

• `docs/mcp-integration/*

Frameworks plugin

@dtbuchholz to review and simplify any here:

• docs/framework-guides/index.json
• docs/framework-guides/ai-sdk.json
• docs/framework-guides/openai.json
• docs/framework-guides/langchain.json

@mzkrasner to review and simplify any here:

• docs/framework-guides/eliza.json

@andrewxhill to final review

• docs/framework-guides/mastra.json

Competitions

@mzkrasner please review these, any simplifications/improvements.

• `docs/competitions/*

Maybe get carson and or kristal's eyes on these pages at some point before merge too

Trading sim

@mzkrasner please review these, any simplifications/improvements.

• `docs/trading-simulator/*

Advanced development

This is old content, is any of it still useful?

• docs/agents/plugins/eliza/cot.mdx
• docs/agents/plugins/eliza/index.mdx
• docs/agents/plugins/eliza/s3.mdx
• docs/agents/plugins/eliza/meta.json
• docs/agents/access.mdx
• docs/agents/index.mdx
• docs/agents/meta.json

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
docs	❌ Failed (Inspect)			Apr 2, 2025 6:17am

[dtbuchholz]

@mzkrasner btw, i can do a first pass at all of the trading sim / other tools you're building since i've been a consumer of them. so i can tag you thereafter for a final pass.

[rohhan]

@andrewxhill

1. I think we received some feedback from Danny and others that "credibly neutral" hasn't historically landed or done well with audiences, so we should move away from that language. It also might not be fully accurate/raise unnecessary questions in early phases (and maybe not even in later phases if we explore subjective competitions). How do you feel about removing that type of language?
2. Would it make more sense to push the Competition Schedule page a bit later once we have more details?
3. IMO we should wait to publish the "Evaluation metrics" section until we have more clarity internally cc @KGmajor

[andrewxhill]

@rohhan

I think we received some feedback from Danny and others that "credibly neutral" hasn't historically landed or done well with audiences, so we should move away from that language. It also might not be fully accurate/raise unnecessary questions in early phases (and maybe not even in later phases if we explore subjective competitions). How do you feel about removing that type of language?

go for it, what's the edit you recommend?

Would it make more sense to push the Competition Schedule page a bit later once we have more details?

Yes, that was my thinking. I wanted to have it drafted, designed in flow wise, and nearly there so that we could touch it up and make it live as soon as we are ready. Versus getting stuck doing another doc overhaul in a week.

IMO we should wait to publish the "Evaluation metrics" section until we have more clarity internally cc @KGmajor

I'd include that in the bucket above, we can remove it from production before we publish. Could you give us a docs timeline here/in notion so we know when these elements should be rolled out?

[dtbuchholz]

Could you give us a docs timeline here/in notion so we know when these elements should be rolled out?

I'll have this stuff ready for Monday. Assuming ppl want to give it a final review thereafter, probably Tues/Weds?

[andrewxhill]

Oh, I meant rohhan for when the different competition elements should go live

…

On Sat, Apr 5, 2025 at 11:18 AM Dan Buchholz ***@***.***> wrote: Could you give us a docs timeline here/in notion so we know when these elements should be rolled out? I'll have this stuff ready for Monday. Assuming ppl want to give it a final review thereafter, probably Tues/Weds? — Reply to this email directly, view it on GitHub <#47 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC2MUZD7IGQ4FR45YPFH632YANALAVCNFSM6AAAAAB2IUZLA2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBRGAZDOOBSGE> . You are receiving this because you were mentioned.Message ID: ***@***.***> [image: dtbuchholz]*dtbuchholz* left a comment (recallnet/docs#47) <#47 (comment)> Could you give us a docs timeline here/in notion so we know when these elements should be rolled out? I'll have this stuff ready for Monday. Assuming ppl want to give it a final review thereafter, probably Tues/Weds? — Reply to this email directly, view it on GitHub <#47 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAC2MUZD7IGQ4FR45YPFH632YANALAVCNFSM6AAAAAB2IUZLA2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOOBRGAZDOOBSGE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[dtbuchholz]

Ooh i misread that ha.

Yeah fwiw, it'd be nice if we rolled out the updates all at once...it's less rework for removing/adding back links to competition pages and whatnot. But only if the competition timelines aren't too far out. Otherwise, nbd.

[dtbuchholz]

closing as #48 incorporates this work