Selora AI: Embedded Claw

Overview

Replace the current web chat in the Selora AI integration with a core Claw layer. This enables users to interact with Selora AI across multiple channels — starting with the existing web chat and Telegram, with plans to add WhatsApp, Discord, and more over time.

We are currently exploring ZeroClaw as a lighter and more secure alternative to OpenClaw for this embedded use case.

Customer value

• Multi-channel access: Talk to Selora AI from the web chat, Telegram, and future channels from a single integration.
• Lightweight and secure: ZeroClaw provides core chat connectivity without the overhead of a full OpenClaw stack.
• Extensibility: New channels can be added incrementally without rearchitecting the integration.
• Flexibility: Advanced users retain the ability to connect their own OpenClaw instances for full feature parity.

Scope (first iterations)

• Core Claw layer: Replace the current web chat backend with an embedded Claw (exploring ZeroClaw as the primary candidate).
• Web chat migration: Migrate the existing web chat to use the new Claw layer.
• Telegram integration: Add Telegram as the first additional channel.
• Channel expansion: Planned support for WhatsApp, Discord, and other messaging platforms.
• External connection: Interface for advanced users to link their own OpenClaw instances.

Architecture (directional)

• Selora AI acts as the orchestration layer.
• An embedded Claw (ZeroClaw) handles messaging between Selora AI and channel connectors.
• Modular channel connectors (web chat and Telegram first, followed by others).
• Secure credential storage for chat tokens and API keys.

Open questions and risks

• Claw selection: Finalizing the embedded Claw engine (ZeroClaw vs. alternatives) based on security and performance audits.
• Web chat migration: Ensuring a seamless transition from the current web chat to the new Claw-backed implementation.
• Feature parity: Determining the exact subset of features to include in the lightweight implementation.

Related

• Selora Smart Agent: Docs