▶ Watch ↗AI Engineer World's Fair 202616:27
Conference affiliation: Inth · 2026
Christopher Burns founded Inth, which builds privacy compliance infrastructure for software teams, and created c15t, its open-source consent SDK. His work connects consent choices and privacy policies to application behavior: which scripts run, how data flows through a product, and whether changes introduce privacy problems.
Burns began building businesses with Will De Ath after university. Their agency work included experiments with NFC and QR codes; a project for a local UK charity led them to co-found Everfund, a donation platform where Burns served as chief executive and lead developer. Everfund combined donation links, embeddable checkouts, and charity dashboards. Its infrastructure handled payment servers, recurring donations, and regulatory requirements so agencies could build fundraising experiences without recreating those systems for every client.
His approach to donation software accommodated different levels of technical control. A small charity could use a ready-made checkout, while a larger organization could build its own interface on the same infrastructure. Understandable tools for nonprofit staff and flexible building blocks for developers were complementary parts of the product.
Alongside Everfund, Burns co-hosted FSJam, a podcast exploring full-stack JavaScript, developer tools, and their builders from 2020 to 2023. He also built its TypeScript-powered website.
Everfund closed its service following a November 2024 notice. In his farewell essay, Burns reflected on seven years building software and a business, including supporting a local food bank during COVID. He considered taking an engineering role, but a privacy problem he had encountered at Everfund drew him back into founding: the relationship between consent systems, privacy policies, and trust in the software handling donations.
His second company began as Consent.io. By spring 2025, he had released c15t for developers who wanted consent infrastructure they could compose and inspect within their application stack. His founding principles emphasized open source, performance, granular control, and visibility into consent state. A cookie banner was the first workflow to improve; the work also addressed how an application enforces the choices collected through it.
c15t provides prebuilt interfaces and headless APIs, controls analytics and advertising scripts according to consent, and supports hosted or self-hosted backends for consent records and audit logs. Those pieces let developers connect the interface to the behavior it promises. Showing a banner is only one step; the application must also respect the person's choice.
Burns renamed Consent.io to Inth on April 13, 2026, reflecting an expansion into privacy policies, data requests, vendor oversight, and compliance automation. Inth joined Y Combinator's Spring 2026 batch. Its platform audits code, pull requests, and websites for privacy issues, including new data flows and discrepancies between loaded scripts, consent state, and policies.
Burns's privacy engineering approach follows product changes through their consequences. Adding an analytics SDK, for example, can change data handling and require consent controls, policy updates, and subsequent review. He advocates improving one workflow at a time while connecting it to the wider compliance process. Performance is another concern: he created Cookie Bench, an open-source benchmarking tool for examining how third-party components and services affect web performance.
As coding agents began recommending and installing libraries, Burns extended his focus on developer experience to the documentation those agents consume. In his 2026 AI Engineer talk, he reported that recommendations from AI assistants had become c15t's largest inbound source, based on onboarding responses. He described several practical changes behind that work while cautioning that the field was changing quickly and his observations were not definitive scientific findings.
.md URLs, requests accepting Markdown, and a query parameter for agents unable to set request headers. He also described early documentation tools for searching, retrieving pages, and asking questions through WebMCP, while acknowledging uncertainty about agent support for some discovery mechanisms.AGENTS.md file directing agents to it. That gives an agent working in node_modules a local route to documentation shipped with the installed library. He reported token savings approaching 50 percent across several models compared with finding documentation through web searches; the talk did not establish that result as a universal benchmark.Burns applies a similar attention to interaction in his presentations. He built and open-sourced shadcn-deck, a Next.js presentation system with reusable React slides, presenter tools, and support for embedding interactive product components. His view of presentations as products treats a demo as another opportunity to show the performance and behavior of the software itself.
▶ Watch ↗AI Engineer World's Fair 202616:27
Loading more talks and highlights…