We’re joining Y Combinator
OpenTag is part of Y Combinator’s Summer 2026 batch. YC is backing us as we build a model-agnostic AI coworker that lives where the work already happens.
This is the second company our three-person team has built together. We worked together at Lilac Labs (YC S24), and OpenTag grew out of a problem we felt there every day: too much company work and context depended on the people already carrying all of it.
What Y Combinator is
Y Combinator is an early-stage investor and startup accelerator. Its three-month program brings each batch of founders together for funding, direct advice from YC partners, focused time with other founders, and a Demo Day where companies present what they have built.
Some of the best-known companies YC has backed include Airbnb, Stripe, DoorDash, Coinbase, Instacart, and Reddit. They span travel, financial infrastructure, local commerce, crypto, groceries, and online communities, but each went through YC while it was still building the foundation of what it would become.
The coworker we wanted
Important work rarely begins in an empty prompt box. It begins in a channel, surrounded by the decisions, people, and history that give the request meaning. Moving that job somewhere else means rebuilding the context before any useful work can start.
With OpenTag, you tag an AI coworker in Slack, give it a job, and get the result back in the thread where the whole team can see it. It can gather information from approved tools, work on its own computer, ask for help when it is blocked, and stop for approval before anything irreversible.
Why model-agnostic matters
The tag is a familiar interface. The durable product is the multiplayer harness underneath it: the context, tools, permissions, memory, and controls that turn a model into a coworker a team can actually use.
Models are improving too quickly to make one provider a permanent product decision. OpenTag can route each step to the model best suited to it, so the team keeps the same workflow while benefiting from better performance and lower costs as the underlying models change.
Context should compound
As OpenTag works, it writes down how the company actually runs: how decisions get made, who owns what, and which workflows keep repeating. That living wiki becomes shared context for teammates and agents instead of another handbook someone has to remember to update.
The boundary stays explicit. OpenTag reads only the channels it has been invited to, uses the connections the team approves, and returns its work with sources. Company memory is useful only when access and accountability grow with it.
What comes next
During S26, we’re staying close to our early teams and focusing on the unglamorous details that make delegation reliable: better context, safer action, clearer results, and more useful work completed end to end.
We’re grateful to Y Combinator, to the teams who trusted OpenTag early, and to everyone who has challenged the idea and helped sharpen it. We’re just getting started.
From Y Combinator
Read the official links
One line to keep
“An AI coworker should compound your company’s context without locking your team to a single model.”
Published July 5, 2026
Back to the blog