Overview
After Tokens 2026 is a one-day, in-person conference in San Francisco about the infrastructure needed to build and operate AI products beyond the model call. Hosted by Stigg, the event focuses on how AI usage is measured, controlled, packaged, and paid for. Its central idea is that token counts alone do not explain what a customer is entitled to use, when an AI agent should stop, or how a commercial contract becomes an enforceable limit in software.
The conference is intended for engineers building AI products and the systems behind them, including backend, platform, and AI infrastructure teams. Technical founders and engineering leaders are also welcome. The page emphasizes practical usefulness to people writing and operating code. Attendees do not need to use Stigg: the event is described as a cross-team and cross-stack conversation, not a Stigg user conference.
Topics and format
The event will cover metering, credits, entitlements, pricing, billing, and the systems that keep these components in sync. The announced format includes technical talks, architecture breakdowns, and lessons from systems running in production. The page does not yet provide a timed schedule, session titles, confirmed speakers, or a venue; speakers, venue, and schedule are still to be announced.
The call for proposals invites practitioners to share a system they built, a failure or incident, and what they learned. Suggested questions include how to meter an agent run, what changed when moving from tokens to credits, and how to check entitlements without slowing down a request. Proposals should explain architecture, implementation, and tradeoffs, and explore a focused engineering problem in depth. Migration stories, production incidents, and unexpected edge cases are welcome. The organizers specifically discourage product pitches and generic talks about the future of AI.
Date and attendance
After Tokens takes place on November 10, 2026, in San Francisco. It is a one-day, in-person event. Attendees can register through Luma, and prospective speakers can submit proposals through Sessionize. The venue and schedule will be announced later. The page does not describe an online attendance option.
FAQ
Who should attend? Engineers building AI products and infrastructure, including backend, platform, and AI infrastructure teams, as well as technical founders and engineering leaders. The talks are meant to be useful to engineers working directly with the code.
Is this a Stigg user conference? No. Stigg hosts the gathering, but participation and the conference conversation are not limited to Stigg customers or users.
How can I attend or propose a talk? Register through the event’s Luma page. To propose a talk, use the After Tokens submission page on Sessionize and focus on a real system and the engineering decisions behind it.
The conference is presented as a place for technical practitioners to exchange concrete approaches to the operational and commercial systems surrounding AI products, with production experience and implementation details at the center.
