Speakeasy

1K posts

Speakeasy banner
Speakeasy

Speakeasy

@speakeasydev

Power your agents with all the tools they need. MCP, CLIs, SDKs and more.

SF & London 🍻 Bergabung Mayıs 2022
70 Mengikuti1.4K Pengikut
Speakeasy
Speakeasy@speakeasydev·
That's the difference between an MCP pilot and an MCP rollout.
English
0
0
0
18
Speakeasy
Speakeasy@speakeasydev·
No configuration per user. No manual provisioning per team. No tickets to IT. Just deploy, and your org has access.
English
1
0
0
25
Speakeasy
Speakeasy@speakeasydev·
The hardest part of rolling out MCP across an org isn't building the servers. It's everything that happens after. 🧵
English
1
0
0
79
Speakeasy me-retweet
Brian Scanlan
Brian Scanlan@brian_scanlan·
We've been building an internal Claude Code plugin system at Intercom with 13 plugins, 100+ skills, and hooks that turn Claude into a full-stack engineering platform. Lots done, more to do. Here's a thread of some highlights.
English
78
196
3K
787.8K
Speakeasy
Speakeasy@speakeasydev·
If you're building MCP servers, trying to run them in production, or thinking about what it takes to make your whole org work through Claude, we'd love to talk. speakeasy.com/mcp
English
0
0
0
24
Speakeasy
Speakeasy@speakeasydev·
The infrastructure for making your entire organization AI-native exists today. The bottleneck was never the AI itself.
English
1
0
0
23
Speakeasy
Speakeasy@speakeasydev·
One week since we launched Speakeasy MCP Platform. Here's what we've seen. 🧵
English
1
0
0
99
Speakeasy
Speakeasy@speakeasydev·
And when things go wrong: error categorization, version correlation, and root cause signals. Not just a dashboard that tells you something broke. Context that helps you understand why and fix it.
English
0
0
0
20
Speakeasy
Speakeasy@speakeasydev·
Latency, error rates, and usage patterns broken down by individual tool. Understand adoption across your org and catch bottlenecks before they become incidents.
English
1
0
0
30
Speakeasy
Speakeasy@speakeasydev·
Once MCP servers hit production, most teams are flying blind. 🧵 They can see that tool calls are happening. They can't see why they're failing, which version handled a request, or whether agents are using tools the way they intended.
English
1
0
0
123
Speakeasy
Speakeasy@speakeasydev·
This is what AI-native looks like in practice.
English
0
0
0
28
Speakeasy
Speakeasy@speakeasydev·
"I don't want anyone using internal UIs or opening SaaS apps. The moment someone shifts a workflow to being AI-native it drops from a two-hour task down to minutes." Shreyas Kumar, Co-founder, Fermat
English
2
0
0
52
Speakeasy
Speakeasy@speakeasydev·
Fermat made Claude the operating system for their entire company. 100% of employees on Claude. 2 days to rollout. 500% MoM growth in AI usage. 🧵
English
1
0
0
120
Speakeasy me-retweet
Rhys
Rhys@RhysSullivan·
Ok last MCP tweet I'll do for now: Imagine if we started with OpenAPI instead - would you be annoyed at people for having giant OpenAPI specs being dumped into context? No that's crazy, you'd be happy to have all the options but you'd want a better way to call them
English
25
3
120
12.5K
Speakeasy me-retweet
Rhys
Rhys@RhysSullivan·
the overindexing on CLIs is kind of insane to me it's building a primitive that's not portable, properly discoverable, has no good approval flow DCR / CIMD to APIs would go so much further but CLI is just the current hype thing
English
45
15
351
52.9K