Floword Workflow Backend

Example Use Cases

• 1

  Orchestrate multi-step workflows described in natural language and execute them against external tools or services.

• 2

  A backend with configurable LLM providers, system prompts, and model kwargs to tune conversational and workflow behavior.

• 3

  Stream execution events in distributed setups (Redis) and register external MCP tool servers via a central mcp.json configuration.

Description

Floword is a backend implementation that helps developers construct and run workflows described in natural language. It provides configuration via environment variables, supports JWT authentication, multiple database backends (SQLite/Postgres), Redis-based streaming, and pluggable MCP tool server definitions via a JSON config. It includes sensible defaults for LLM model providers and system prompts, Docker support, and development tooling to accelerate building NLP-driven automation and integrations.