AgentEra/Agently

[GenAI Application Development Framework] 🚀 Build GenAI application quick and easy 💬 Easy to interact with GenAI agent in code using structure data and chained-calls syntax 🧩 Use Event-Driven Flow *TriggerFlow* to manage complex GenAI working logic 🔀 Switch to any model without rewrite application code

What it solves

Agently is an AI application runtime framework designed to move AI services from experimental prompts to reliable production backends. It solves the problem of "output drift" and model instability by providing stable contracts for structured outputs, observable action execution, and recoverable workflows, ensuring that business logic remains intact even when the model or provider is switched.

How it works

Agently normalizes the interaction between the application and the AI model through a unified request/runtime contract. It uses "prompt slots" to compose requests and a contract-first approach to structured output, where schemas are defined as prompts and validated by the framework. For orchestration, it includes TriggerFlow, a signal-driven workflow engine that allows for pause/resume capabilities and event-based execution. It also manages a dedicated Action Runtime for tool use, supporting local functions, MCP servers, and sandboxed execution environments (like gVisor or Landlock).

Who it’s for

  • Product Engineers building AI assistants, internal copilots, or AI-backed APIs who need reliability and structured data.
  • Platform Teams requiring clear extension points for model providers, tools, and observability.
  • Technical Leads looking for a maintainable framework that prioritizes explicit control and debuggability over "prompt glue."

Highlights

  • Contract-First Output: Guarantees structured output through schema-as-prompt, required-field extraction, and validation handlers.
  • Structured Streaming: The instant mode allows applications to react to specific structured fields while the model is still streaming.
  • Observable Actions: Provides a managed Action Runtime for tools, including built-in support for Python, Shell, Node.js, SQLite, and MCP servers.
  • Signal-Driven Workflows: TriggerFlow supports event-driven DAGs, fan-out, and state snapshots for recoverable execution.
  • Model Agnostic: Normalizes provider setup and response parsing, making it easy to switch models without rewriting business logic.

Related

  • Project
    alikon-art/DeterminFlow

    DeterminFlow is an open‑source, AGPL‑licensed framework for building, running, and monitoring production‑grade AI workflows. It breaks complex pipelines into isolated LLM‑Agent nodes, scripts, API calls and human approvals, providing versioning, checkpoint‑based recovery, per‑node token accounting, and a visual editor. The tool is aimed at content generation, data‑centric AI pipelines, and enterprise automation, and can be installed via desktop binaries, run from source, or deployed with Docker.

  • Project
    agentscope-ai/agentscope

    A production-ready agent framework that provides an SDK and a FastAPI-based service to build, deploy and scale multi-agent applications with built-in RAG and tool management.

  • Project
    Eigenwise/atomic-agents

    A modular framework for building predictable agentic AI pipelines using Pydantic and Instructor to ensure consistent, structured outputs.

  • Project
    VRSEN/agency-swarm

    A framework for building multi-agent AI applications that organizes agents into collaborative swarms based on real-world organizational structures.

  • Project
    agentspan-ai/agentspan

    A durable runtime for AI agents that prevents execution loss during process crashes and supports long-running, event-driven workflows with human-in-the-loop approvals.