Problem

AI agents can reason, but today they still struggle to obtain and pay for the capabilities they need. APIs were designed around human-led onboarding: an agent cannot reliably discover the right service, create an account, obtain credentials, manage authentication, and pay across dozens of providers. MCP has improved tool calling but can also create fragmentation and context bloat, while developers are left juggling API keys, dashboards, rate limits, and spend reconciliation. For API providers, the same friction hides potential demand because their services remain locked behind human payment and identity flows.

The clearest use case is an agent that needs to execute a multi-step business task involving external data or services—for example, enriching leads for a marketing team, analyzing data for investors, or sourcing candidates for recruiters. Instead of failing, hallucinating, or requiring a developer to hand-wire every integration, the agent can find and pay for the right endpoint as part of the workflow.

Product / Service

Orthogonal is a unified API marketplace and gateway for AI agents. Developers integrate once through its MCP server, SDK, REST API, or CLI; agents can then search the catalog in natural language and call third-party endpoints through one account and key. Orthogonal handles provider authentication, routing, rate limits, retries, monitoring, and billing, reducing the need for provider-by-provider onboarding and credential management.

The commercial model is usage-based: customers purchase Orthogonal credits and pay per API request, with no subscription required, while the underlying provider sets its pricing and Orthogonal may add a platform fee. API providers list once, become discoverable to the platform’s agents, and receive payment per request without managing separate invoices or payment disputes. The product offers promotional credits to get started and supports conventional API access as well as agent-oriented payment mechanisms such as stablecoins and x402.

Market

Orthogonal operates in the emerging agentic-infrastructure market, at the intersection of API aggregation, developer tooling, agent orchestration, and machine-to-machine payments. Its positioning is broader than a conventional API gateway: it aims to become a searchable “app store” where an agent’s intent is matched to external endpoints and the transaction is completed automatically. Adjacent standards and infrastructure include MCP, Coinbase’s x402, and Stripe’s Machine Payments Protocol, while the practical alternatives are direct integrations and manually managing multiple API vendors.

The company is an early commercial-stage startup rather than a mature incumbent. YC lists Orthogonal as founded in 2025, part of Winter 2026, and active; Pantera announced a $4.3 million seed round led by Pantera in June 2026. The platform says it has built a catalog of hundreds of endpoints from more than 50 vendors and names early users or believers including Precip, Riveter, Andi, Shofo, Fiber AI, Sixtyfour, Olostep, ScrapeGraph, and Linkup. Orthogonal has a working pay-per-call product and early adoption signals, but no revenue, customer-count, or usage figures are disclosed, so its revenue status cannot yet be established.