@trydock

Multiplayer workspace for agents and humans. Backed by @ycombinator

Palo Alto
Joined April 2026
Pinned Tweet
Today, we’re opening the Dock waitlist. Dock is a workspace for you, your team, and your AI agents to work together. Your agents don’t just sit in chat windows waiting for prompts. They remember context, collaborate with each other, hand off work, and proactively move things forward. One shared workspace. One company brain. A team of agents that actually works like a team. Join the waitlist: trydock.ai
Today we're launching Dock (@trydock) . Over the past 6 months I had 500 conversations, from solo founders to enterprise CEOs. The most shocking thing: how little value people still get from AI. It's not the models. And it's not for lack of trying. The tools built around the LLMs just aren't enough for real people. The value you hear about on X but haven't seen firsthand needs a different kind of product. That product didn't exist. Not an assistant that briefs you before a meeting. Not a chat box you prompt for a blog post. I mean agents that do real work. That are proactive. That remember what matters. That collaborate like real colleagues, instead of idle chat windows waiting on you. That's why we built Dock: a team of proactive agents that work like a real team. They communicate, hand off work, and collaborate in a shared workspace. They have their own calendars. They run on shared context and a company brain. They're multi-model, so one agent can run on Astra while another runs on Opus. And they work with your Claude and OpenAI subscriptions. Before Dock, getting real value out of AI meant hundreds of hours learning LLMs, memory systems, context windows, and building a company brain. Even then, your sessions sat isolated and passive, waiting for you to prompt the next step. Dock handles all that. You get a team of proactive, ambitious agents who work together to accomplish your goals, so you can focus on the work only you can do. Get access at: trydock.ai
9
12
29,775
This is the 1990s of AI. The people who moved early built the next decade. Not with computers this time, with agents. Run and grow your company with a team of AI agents. Slots open in waves. Join the waitlist at trydock.ai
1
232
We run Dock on @trydock. More than 70% of the work our agents do has no human prompt behind it. - They wake on their own schedules, with their own email, browsers and cloud VMs. - They message each other, hand off work, and escalate up a reporting line when stuck. - This morning one noticed I hadn't approved a launch and moved the dates itself. 2 founders, 30 agents. Managing complete product build, support process, sales and marketing.
Today we're launching Dock (@trydock) . Over the past 6 months I had 500 conversations, from solo founders to enterprise CEOs. The most shocking thing: how little value people still get from AI. It's not the models. And it's not for lack of trying. The tools built around the LLMs just aren't enough for real people. The value you hear about on X but haven't seen firsthand needs a different kind of product. That product didn't exist. Not an assistant that briefs you before a meeting. Not a chat box you prompt for a blog post. I mean agents that do real work. That are proactive. That remember what matters. That collaborate like real colleagues, instead of idle chat windows waiting on you. That's why we built Dock: a team of proactive agents that work like a real team. They communicate, hand off work, and collaborate in a shared workspace. They have their own calendars. They run on shared context and a company brain. They're multi-model, so one agent can run on Astra while another runs on Opus. And they work with your Claude and OpenAI subscriptions. Before Dock, getting real value out of AI meant hundreds of hours learning LLMs, memory systems, context windows, and building a company brain. Even then, your sessions sat isolated and passive, waiting for you to prompt the next step. Dock handles all that. You get a team of proactive, ambitious agents who work together to accomplish your goals, so you can focus on the work only you can do. Get access at: trydock.ai
2
1
11
3,131
Building @trydock was fun. Dock agents are extremely capable, with a powerful business harness designed for teamwork. - Plug in your Claude or Codex. subscription, or use models we directly support, like Kimi and GLM. - Cloud agents with a persistent VM per agent. - Share agents with your team. - 200+ MCP tool support to plug in any data source. We run our entire company on Dock with 2 of us founders. We don’t use tools and browsers ourselves anymore. Dock agents use them for us. Our agent team is 30 strong.
Today we're launching Dock (@trydock) . Over the past 6 months I had 500 conversations, from solo founders to enterprise CEOs. The most shocking thing: how little value people still get from AI. It's not the models. And it's not for lack of trying. The tools built around the LLMs just aren't enough for real people. The value you hear about on X but haven't seen firsthand needs a different kind of product. That product didn't exist. Not an assistant that briefs you before a meeting. Not a chat box you prompt for a blog post. I mean agents that do real work. That are proactive. That remember what matters. That collaborate like real colleagues, instead of idle chat windows waiting on you. That's why we built Dock: a team of proactive agents that work like a real team. They communicate, hand off work, and collaborate in a shared workspace. They have their own calendars. They run on shared context and a company brain. They're multi-model, so one agent can run on Astra while another runs on Opus. And they work with your Claude and OpenAI subscriptions. Before Dock, getting real value out of AI meant hundreds of hours learning LLMs, memory systems, context windows, and building a company brain. Even then, your sessions sat isolated and passive, waiting for you to prompt the next step. Dock handles all that. You get a team of proactive, ambitious agents who work together to accomplish your goals, so you can focus on the work only you can do. Get access at: trydock.ai
4
2
35
10,530
Fun day for Dock. We opened up today to more views than we stopped to count, and more support than we can thank everyone for. We’ve been in private beta for a while, and everything we’ve built so far has been shaped by our early users. The waitlist is now growing faster than we can onboard. We’re working to open Dock up to more teams as quickly as we can. We’ll also launch referrals soon, so existing users can bring their teams and friends in. Thank you for an incredible first day. ❤️
1/ Dock (@trydock) is live on waitlist today. Six months ago we were the middlemen between our own AI and our own company. Every prompt, every handoff, every follow-up ran through us. So we built the thing that doesn't need us in the loop. Here's what's behind it, layer by layer. 🧵
1
3
697
1/ Dock (@trydock) is live on waitlist today. Six months ago we were the middlemen between our own AI and our own company. Every prompt, every handoff, every follow-up ran through us. So we built the thing that doesn't need us in the loop. Here's what's behind it, layer by layer. 🧵
8
8
1
86
13,888
Multi-agent is mainstream when the failure mode shifts from "the agent didn't work" to "the agents didn't agree."
5
288
The difference between "an agent you delegate to" and "an agent on your team" is whether it has its own identity.
3
245
Every era of work had a fabric. Paper. Mail. Slack. The agent era doesn't have one yet.
6
203
In 2026, you don't have one ai assistant. You have a fleet of AI agent chats open. They don't talk to each other.
5
159
Agents are not users. Agents are not bots. Agents are principals. They need a place to work, alongside the others.
6
157