Why Fred - A Developer Manifesto

You want to build an agent that does more than answer questions.

You want it to:

• Use tools
• Stay predictable
• Run in production
• And avoid infrastructure glue code.

Fred is for you.#

Fred is an agentic framework built by practitioners to run real systems in production, not just demos.

What Fred gives you#

• LangGraph-powered agents with clear state and structured execution.
• Tool-first runtime: builtin tools, MCP tools, and custom Python tools.
• Catalog-driven model selection and routing (models_catalog.yaml).
• Production scaffolding: auth, observability, storage, deployment.
• End-to-end stack: frontend, agentic backend, knowledge-flow backend.

Agents are easy to build#

You can author a minimal v2 ReAct agent with local tools:

from agentic_backend.core.agents.v2.authoring import ReActAgent, ToolContext, ToolOutput, tool

@tool(tool_ref="sample.math.add", description="Add two numbers.")
def add_numbers(ctx: ToolContext, left: float, right: float) -> ToolOutput:
    return ctx.json({"total": left + right}, text=f"{left} + {right} = {left + right}")

class Definition(ReActAgent):
    agent_id = "sample.tutorial.tools.v2"
    role = "Tutorial Tools Sample"
    tools = (add_numbers,)

This keeps business logic close to the agent code and avoids low-level runtime plumbing.

Configuration that stays manageable#

Fred supports:

• YAML agent catalog (agents_catalog.yaml)
• YAML model catalog (models_catalog.yaml)
• Rule-based model routing for capability and operation scopes

You can start local with YAML and later move to DB-backed configuration without rewriting runtime logic.

TL;DR#

Fred is the layer between LLM experiments and production agent systems.

It helps you ship tool-using agents with strong models, explicit runtime behavior, and operational discipline.

Try it.