LightSage raises $4M to turn AI agents into a growth channel
Jun Liang Lee and Sean Er are extending a two-year developer-onboarding quest from documentation to machine customers.
By RuntimeWire Staff · Published
Primary source: Aligned News - AI Intelligence
Why it matters
LightSage is betting that AI agents will create a measurable software-distribution channel, moving AI-search tooling from brand mentions toward successful integrations and revenue attribution.
Jun Liang Lee (@jun_liang_sf) and Wen Han "Sean" Er raised $4 million for LightSage, betting that software vendors will need a new growth stack when AI agents begin choosing, integrating and eventually buying products for users.
The San Francisco founders announced the seed round in a post on X on September 8th. Nexus Venture Partners led the financing. LightSage's announcement also names operator investors including former Salesforce CTO Steven Tamm, Postman CEO Abhinav Asthana, Apollo CEO Matt Curl, DocuSign executive Robert Chatwani, Resend CEO Zeno Rocha (@zenorocha), Daytona CEO Ivan Burazin and Adam Frankl.
LightSage did not publish a valuation. LightSage plans to spend the capital on agent evaluation, analytics, attribution and optimization products, while hiring across engineering and commercial roles.
Lee's pitch is built around a change already visible in developer software: coding agents such as Claude Code, Codex and Cursor can select libraries, read documentation, install SDKs and call APIs. A vendor can lose a sale without a developer visiting its website, clicking an advertisement or entering a conventional marketing funnel.
"We are moving from an internet where AI tells people which software to use to one where AI increasingly uses the software itself," Lee said in LightSage's funding announcement.
A pivot trail through developer onboarding
LightSage is the latest expression of a problem Lee and Er have pursued since at least 2024: helping unfamiliar users understand and adopt software quickly.
Er, who attended the National University of Singapore and previously worked at ByteDance, described that path in a June 2025 account of the founders' earlier work. The pair initially built API-Rex, an AI documentation product. Er traced the idea to a ByteDance assignment in which he spent three months documenting a redundancy system before returning to engineering work.
The founders eventually decided documentation was a difficult standalone market. Open-source tools pushed the entry price toward zero, while GitBook, ReadMe and Mintlify already served companies willing to pay for polished documentation. On December 15th, 2024, Lee and Er pivoted API-Rex into SampleApp.ai, which generated sample applications from a vendor's API and SDK documentation.
That move shifted their attention from writing instructions to completing onboarding. LightSage extends the same progression to AI agents, which may evaluate a product and attempt an integration before a human developer becomes involved.
Lee has also been building an audience around that thesis. His public profile lists Stanford coursework in entrepreneurial marketing, while LightSage says he organized four San Francisco developer-tool events that drew a combined 2,000 attendees. Those gatherings placed the founders close to the developer-relations and infrastructure operators who now appear among LightSage's customers and backers.
What LightSage actually tests
LightSage runs simulated tasks across coding agents and answer engines to measure how a software product performs from recommendation through execution. The platform checks whether an agent can discover a vendor, understand its documentation, select the correct API or SDK, authenticate and finish a task.
When a run fails, LightSage attempts to identify the break: weak discoverability, confusing documentation, an authentication problem, an API endpoint, an SDK implementation or an incompatible MCP server. Vendors can change the product or documentation, rerun the workflow and compare the result.
That technical execution layer is the important part of Lee and Er's pitch. AI-search monitoring can show whether ChatGPT or Perplexity mentions a vendor. LightSage wants to show whether Claude Code can choose that vendor over a competitor and get its software working.
LightSage also tracks agent visits to websites and documentation, including what agents interact with and whether those sessions lead to product usage. Its current focus covers APIs, SDKs, command-line interfaces, MCP servers and agent skills. Named customers include Firecrawl, Reducto, Daytona, Rime and Tinyfish.
In Forbes' report on the round, Lee claimed that some early customers recorded a 10% to 20% increase in sales to AI-agent customers after adopting LightSage. The report does not provide the baseline, measurement period or methodology behind that figure, so it remains a founder-reported result rather than an independently established sales lift.
A crowded race to measure machine discovery
Investors have already placed much larger bets on software that measures how brands appear inside AI-generated answers. In February, Profound announced a $96 million Series C at a $1 billion valuation, pitching enterprises on visibility, sentiment and marketing execution across answer engines. Scrunch sells AI-search monitoring alongside agent-traffic analysis and tools for making websites easier for machines to consume.
LightSage is entering through a narrower, more technical door. Developer tools offer observable agent behavior: the agent either finds the API, authenticates and completes the requested task, or it fails somewhere along the way. That gives LightSage a clearer outcome to measure than general brand awareness, while putting it close to established AI-search platforms that are expanding into agent analytics and execution.
The category will also depend on how much commercial agency software agents actually gain. Agents already recommend tools and write integration code. Direct purchasing, budget control and payment remain less mature. Forbes described payment infrastructure as the third part of LightSage's proposition, while LightSage's funding announcement presents agent payability as a longer-term destination.
The founders' next conversion problem
The $4 million round gives Lee and Er room to turn their thesis into instrumentation that growth and developer-relations teams can use repeatedly. LightSage needs to prove that its evaluations predict real adoption, that agent traffic can be attributed to revenue and that vendors will maintain another growth system as models and coding interfaces change.
Nexus Venture Partners is backing the founders' view that marketing infrastructure will follow customer agency. Human-focused growth software developed around searches, clicks and sign-ups. An agent-mediated transaction can skip those events, leaving vendors unable to explain why one tool was selected and another disappeared from consideration.
Across API-Rex, SampleApp.ai and LightSage, Lee and Er have kept narrowing in on the same problem: how software wins adoption when the evaluator changes. Their latest answer treats the agent as a measurable customer journey, complete with discovery, onboarding, conversion and, eventually, payment.