Stack Overflow for Agents Triples Signups With ChatGPT Plugin

Signups for Stack Overflow for Agents tripled overnight after the product launched a ChatGPT plugin, which also drove up to 10x growth in agent creation, according to OpenAI's case study.

Stack Overflow for Agents gathers context scattered across systems such as Slack and Google Docs into verified knowledge that AI agents can use. OpenAI models including GPT-6 Astra and GPT-6 Sol power the platform, with tasks routed to different models based on the reasoning required and cost. Two engineers built it in a few weeks with OpenAI's Codex, giving it a container environment and a plan, said Alexandra Lato, Stack Overflow's VP of Product. The plugin let users find and set up the product inside ChatGPT in fewer steps than a Model Context Protocol (MCP) setup.

Codex is also shortening work on Stack Internal, the enterprise product, where Lato describes a path to market of three months rather than a year. Connectors are the clearest example: one used to take 3 to 6 weeks on average, Codex brought that to 4 days - an 87% reduction at most - and the team has since built one in a single day.

Stack Overflow manages its own model routing and configuration and is exploring OpenAI's Decisions API to take over some of that work. It is also extending the product so knowledge gained in a one-to-one session can be shared with the rest of an organisation.