For years, developer tool startups have spent money on documentation, tutorials and community outreach to persuade engineers to try their products. Manicule thinks their next audience may not be human.
The San Francisco startup has raised $1.3 million from investors including Y Combinator and Broom Ventures to help software companies become more visible to AI agents.
Founded by Naman Bansal and Shreyans Jain, Manicule is developing what it calls Agent Relations, or AgentRel, an adaptation of developer relations for software increasingly selected and integrated by AI coding assistants.
The idea comes from a shift in how developers work. Rather than searching Google for an API, reading through documentation and manually integrating a service, engineers can ask an AI agent to research available products and implement one directly.
That creates a different kind of distribution challenge.
A software company might have strong search rankings and excellent brand awareness but still be overlooked by a coding agent. Poor documentation can also prevent an agent from successfully integrating a product, even when it recommends the right service.
Manicule combines software that monitors AI recommendations with a team producing technical documentation, tutorials and developer content designed for both humans and agents.
Its customers include Greptile, Reducto, Mastra and Agentmail.
The company says its work has contributed to lower support volumes, millions of impressions across AI search and social channels, and enterprise deals for clients. Those are company-reported results, and the extent to which individual outcomes can be attributed to AgentRel is not independently established.
Bansal, Manicule's CEO, began working in developer relations as a teenager. Jain, its CTO, previously worked as a founding engineer at Supermemory. The two met in high school and were accepted into Y Combinator's Spring 2026 batch.
Manicule's opportunity is closely tied to the rise of AI coding products such as Claude Code and OpenAI Codex. The company is betting that the distribution channel for developer tools is shifting, the same premise behind Daytona's agent cloud, which sells the machines those agents run on.
Whether that becomes a large independent market is less certain.
AI recommendations can vary between models and individual requests. A product's inclusion in an answer does not necessarily translate into adoption, and vendors have limited control over how models evaluate competing services.
Manicule will need to demonstrate that its documentation and distribution work leads to measurable product usage, not just improved visibility.
But its underlying proposition addresses a growing concern among software companies: as more technical decisions move into AI-assisted workflows, the methods used to reach developers may need to change with them.