CTO, Co-Founder, @tryArcade | @redisinc @HPE_Cray | AI, LLM, Python, HPC, ⚾️ fan, roll tide | Blog: https://xitter.cf/t.co/PJfL2cMCYD | Opinions are my own.
San Francisco, CA
Joined April 2013
- Tweets1.4K
- Following463
- Followers1.7K
- Likes4K
Sam Partee retweeted
Look Ma, we made it again! 🎉
I have a new peer reviewed paper on @arxiv: Tool Recommendation for Multi-Step Agents: Finding Every Tool a Task Needs
Written with the @TryArcade crew: @SamPartee, River Stevens, and Valerie Fanelle.
We started out trying to figure out “how do we give agents the whole ~8k tool catalog without the context window or the bill exploding.”
Ended up with a real deployed system with some pretty nice numbers:
• First recommendation covered every required tool in 93% of Codex sessions
• 57% better top-10 coverage than BM25 on ToolRet
• Cut input tokens 17–40% and the bill by up to 40% vs a fixed list of 100 tools (no drop in task completion)
Link to full paper in the comments
Sam Partee retweeted
@SamPartee , River Stevens, @ptdamiba, and Valerie Fanelle put the research behind Arcade.dev’s Tool Recommendation on arXiv. These are the details behind all the token cost reduction and ToolRet scoring. For the nerds: arxiv.org/abs/2610.23707
Sam Partee retweeted
This week in Chicago, Arcade.dev CTO @SamPartee and Sabry Loganathan of @onepeloton talked through governing enterprise agents in production, from identity models and multi-user policies to the lessons they wish they'd known before scaling past the pilot.
We're hoping to get them on a stage together again soon.
Sam Partee retweeted
Thanks to @VentureBeat for the exclusive on Tool Recommendation’s launch. @carlfranzen nailed the problem it solves: “Connecting an agent to more business systems also makes its instructions larger and more expensive.”
venturebeat.com/orchestratio…
Sam Partee retweeted
We're "already fixing that."
@VentureBeat covered five startups building infrastructure for enterprise agents. Our CTO @SamPartee on the three snags agents hit authorization, governance, and reliability and why acting on behalf of real users takes a different security architecture:
venturebeat.com/orchestratio…
Sam Partee retweeted
Agent identity was always the goal for @SmitheryDotAI.
We talked about it in our blog posts about building "Google for a new internet", the new internet being one in which tool calls outpace human clicks.
With @TryArcade being deployed at multiple F100s including big six megabanks (and raising a $60m series A!), they've found the clearest path to enterprise value in the same space.
Thrilled to be joining such a capable team to bring the Smithery developer experience to the enterprise, and the enterprise-grade robustness and security to the Smithery platform :)
@SmitheryDotAI, the leading public registry and hosting platform for MCP, is now part of Arcade.dev.
Smithery made finding and running an MCP server take one click. Pairing that developer experience with the security, reliability, and governance that Arcade is known for, gives enterprises both to lead the next generation of agentic AI.
Smithery co-founder @kamathematic is joining us to keep building it.
More from @TheMostlyGreat: arcade.dev/blog/smithery-joi…
Smithery 🤝 Arcade
So excited for this and to work with @kamathematic
So begins the path to eating the agent stack from the bottom up
@SmitheryDotAI, the leading public registry and hosting platform for MCP, is now part of Arcade.dev.
Smithery made finding and running an MCP server take one click. Pairing that developer experience with the security, reliability, and governance that Arcade is known for, gives enterprises both to lead the next generation of agentic AI.
Smithery co-founder @kamathematic is joining us to keep building it.
More from @TheMostlyGreat: arcade.dev/blog/smithery-joi…
Sam Partee retweeted
July 28 is the largest MCP spec update to date. We’ve covered stateless and auth. Next up: extensions.
Notice that no client is obligated to support any extension and authors can ship breaking changes on their own cadence.
A tool that works in Claude might not work in Cursor.
Read more for the full breakdown: arcade.dev/blog/mcp-extensio…
Sam Partee retweeted
everyone has a dope AI agent demo. almost nobody has one that survives monday traffic....
On July 2nd @TryArcade is joining {AI} in Production along with @inngest, @cursor_ai, @Vapi_AI.
real prod lessons, free food, good people. Hosted at Inngest HQ
rsvp in replies. come find me 👋
If you're at the Bengaluru or Mumbai @MCPDevSummit come through!
⚙️ Building the infrastructure layer for #AI agents? Don’t miss Sam Partee, CTO and co-founder of Arcade.dev, as he takes the keynote stage at #MCPDevSummit Bengaluru, 9-10 June. Sam brings a practical perspective on what it takes to power the next generation of agentic systems.
🎟️ Register today and join us!
bit.ly/4t1Dica
In honor of @LangChain Interrupt today, here's my favorite Agent I made with Fleet + @TryArcade
Don Forgettabouit - a very sarcastic reminder agent
Started as a simple email triage agent I setup in 5 minutes and now it's the first email I open everyday.
Today's gem - "Did you miss the 'Action Required Today' part of that email from 3 days ago?"
Sam Partee retweeted
We just open-sourced Agent Library, a local-first memory layer for AI agents built by @SamPartee.
Your agent's memory should be something you can hold in your hand. A file you own, diff, and roll back.
More on the blog: arcade.dev/blog/agent-librar…
S/o to @torresmateo for pushing me to release this!
agent-library is just as it sounds - a library for an agent.
It started as my personal obsidian connector for Arcade, but has grown into a multi-modal knowledge management system.
Tweet thread on the design and what I learned building it very soon.
your agent's memory shouldn't live on someone else's computer.
today @TryArcade is open-sourcing Agent Library: your agent's entire memory in one SQLite file you can copy, diff, or email to yourself. Apache 2.0. no account. no hosted service.
github.com/arcadeai/agent-li…
MCP and Skills are not mutually exclusive!!! nor does either have a exclusive locale
it's like people forget MCP used to get flak for being "local only".
Skills can use MCP tools. MCP will soon (conjecture) support skills as resources as many clients do today.
@openclaw ... but give it everything!
Wrote a plugin to give OpenClaw / Moltbot / Clawdbot access to all of @TryArcade tools.
Now you can use 100s of services from any of your msg apps i.e. Discord, Telegram, etc. Auth included
👇Me using GCal and Slack from Telegram
PR for anyone who wants to try before it's merged!
github.com/openclaw/openclaw…
Many people love using @TryArcade + Claude or cursor.
I’m seeing sales people make a gateway with hubspot, google docs, Gmail, attio, etc in arcade and doing their job from cursor.
Underrated is Posthog. “Lookup this customers usage and tell me about it before we meet”
Someone needs to build a company around Customer Context Graph.
Collect all the threads – emails, meeting transcripts, slack messages, contracts, deliverables, detail, info, and config – from your customers into context that can be explored and queried by agents.
This info is scattered between CRMs, ticketing systems, note takers, product, landing pages – it's inherently cross platform information. You need a new solution. Kind of how Segment did it trad SaaS apps.
With this context, you can fire up Claude Cowork or similar for ad-hoc work or build extremely powerful agent automation flows. Expose the context as skills, MCP, and file system.
Even better if you build it as open-source with a hosted option so people can take it on-prem as needed. Create a connector ecosystem around it. This will power every single next-gen AI-native full-stack business.
Sort of like the context graph (@ashugarg @JayaGup10 ) that has been discussed recently but I'm thinking something very concrete: "Get me all the context about this particular customer."
A customer-level, cross-system context substrate that agents can explore and act on