What it does

Tock is a Kotlin-based toolkit for building conversational bots across Messenger, WhatsApp, Slack, Alexa, Google Assistant, and roughly a dozen other channels. It bundles an NLP pipeline (pluggable with OpenNLP, Stanford, Rasa), a visual story builder called Tock Studio, and SDKs for Kotlin, Node.js, and Python. Deployment is strictly BYO: Docker, cloud, or on-premise.

The interesting bit

The project treats “open” as infrastructure, not just license. You get the full stack—NLU, dialog management, analytics, connectors—without a vendor-hosted tier nudging you toward paid plans. The Kotlin DSL for modeling conversations is a nice touch for JVM shops who find Rasa’s YAML exhausting.

Key highlights

• Pluggable NLP: swap intent classifiers without rewriting your bot
• Tock Studio: visual IDE for non-devs to build dialog flows
• 15+ built-in channel connectors, plus React/Flutter toolkits for custom frontends
• Multi-language SDKs (Kotlin, Node.js, Python) and a REST API
• Docker-first deployment; no managed service upsell lurking in the docs

Caveats

• The README is thin on architecture details and leans heavily on external documentation
• 605 stars suggests a modest community; long-term maintenance trajectory is unclear
• Python tooling is bolted on (pre-commit hooks, separate setup) rather than first-class

Verdict

Worth evaluating if you need a self-hosted alternative to Dialogflow or Rasa Enterprise and your team already speaks Kotlin. Skip it if you want a fully managed service or a bustling plugin ecosystem.