Problem

AI agents can reason and execute tasks, but they do not naturally have a durable identity or the communication and authentication channels needed to operate in the real world. Inkbox’s founders encountered this when web-automation agents needed confirmation emails and 2FA codes, while contacting an agent required stitching together Gmail, Twilio, ngrok, and other human-oriented services that shared neither identity nor context. The killer use case is an agent that can independently sign up for services, receive verification codes, maintain accounts, and communicate with people over time. The economic pain is integration sprawl: every agent requires separate provisioning and maintenance, and the burden becomes a liability as teams deploy fleets rather than one-off demos.

Product / Service

Inkbox is a developer-facing API and SaaS layer that provisions an identity for each agent. An identity includes a real email inbox and persistent public tunnel, with an optional phone number for calls and SMS/MMS; the broader product also supports iMessage, shared contacts and notes, and a zero-knowledge encrypted vault for credentials, API keys, SSH keys, and TOTP secrets. These resources are organized around the same agent identity, so email, phone, messaging, webhooks, conversation history, and credentials can be used with shared context rather than as disconnected services.

The service is delivered through APIs, SDKs, documentation, command-line tools, and plugins for agent harnesses such as Claude Code, Codex, Hermes, and OpenClaw. Inkbox offers a free tier, a $30-per-month Developer plan, a $200-per-month Startup plan, and custom Enterprise pricing, making the value proposition a simpler way to provision and operate agents than assembling and maintaining multiple communications and credential products separately.

Market

Inkbox competes in AI-agent infrastructure, specifically developer tools for agent identity, communications, reachability, and shared state. Its initial customers are developers and companies building customer-support assistants, outbound-sales agents, automated phone operators, and other agents that must communicate with real people or follow up over time. The main substitutes are not necessarily identical products but the do-it-yourself stack of Gmail or another mail provider, Twilio for voice and messaging, ngrok for public endpoints, and separate password or 2FA tooling. Inkbox’s differentiation is treating those capabilities as one persistent identity with unified context; adjacent channel specialists such as Sendblue also represent competition for individual messaging functions.

The company is early-stage rather than proven at scale: Y Combinator lists it as an active Summer 2026 company founded in 2025 with a three-person team, while Inkbox has publicly launched documentation, plugins, and usage-based subscription tiers. The available evidence does not disclose revenue, customer counts, or production usage, so Inkbox should be described as launch-stage and pre-scale rather than definitively labeled pre-revenue. Its public pricing and active product suggest commercial availability, but meaningful traction remains unreported.