Orchid

Keiki is a platform from Orchid for launching customer-facing AI agents on channels people already use, including iMessage, WhatsApp, SMS, Slack, Telegram, Teams, and email. Teams define an agent's role, behavior, and boundaries once, connect business knowledge and custom actions, then deploy the same agent wherever customers or coworkers communicate, with Keiki handling channel delivery, conversation memory, tool execution, billing, observability, evaluations, browser automation, and sandboxed code execution. Keiki agents can retrieve live information, complete workflows, take payment, remember relevant context, and escalate to a person when policy requires it. The launch opens up infrastructure the Orchid team spent close to a year building for its own iMessage assistant, now repackaged so any business can ship an agent without stitching together channels, memory, tools, and evals from scratch. Builders can start in a visual agent builder or use the developer interfaces, with Keiki publishing an OpenAPI contract, provider-compatible endpoints, AgentKit documentation, an MCP server, and scoped API-key controls for teams moving from prototype to production. For ongoing operation, teams get conversation history, execution traces, evaluations, usage controls, alerts, and human steering, with the option to require approval before sensitive work proceeds. Keiki comes from the team behind Orchid, founded in 2025 by Nizar Abi Zaher and Adam Wazzan and based in San Francisco , which went through Y Combinator's Spring 2025 batch with a personal assistant that lives inside iMessage. For founders and operators evaluating where to put an agent in front of customers, Keiki is the Orchid team's bet that the right surface is the messaging apps those customers already check every day.