Conversational agents and general agents
Conversational and general-purpose AI agents work via chat, messengers or the desktop and take on changing tasks – from personal assistants to generalists with tool access.
Conversational agents hold conversations in natural language and complete tasks along the way. General agents are not tied to one domain: they combine web search, file access, code execution and other tools to handle different assignments.
Many projects in this category are designed as personal assistants. For use in a company, responsibility, permissions and data access need to be clearly defined.
Role in an AI-first business model
In customer contact, conversations are where revenue is won or lost. Conversational agents can classify enquiries, prepare replies and continue conversations where customers already communicate.
This is exactly where not every reply should be sent autonomously. An AI-first operating model defines which replies an agent gives on its own, which it prepares and when a person takes over.
What to look for when choosing
- Channels: where does the conversation actually take place?
- Handover to people with full context
- Rules for what may be sent without approval
- Data access and storage of conversation content
18 tools in this category
Colossal-AI
A system for training and running large language models. Its ColossalChat application demonstrates how to build a chat model using RLHF (reinforcement learning from human feedback).
RasaGPT
A headless chatbot platform that combines the Rasa dialogue framework with LangChain and LLMs. A starting point for building your own chatbots without a prescribed user interface.
SuperAgent
An open-source project aimed at taking LLM agents into production. Relevant for teams that want to run agents in real use rather than only as prototypes.
ix
A platform for autonomous agents built on GPT-4. It is used to set up agents that work through tasks on their own in multiple steps.
Multi-Modal LangChain agents in Production
A starter project from Steamship for deploying LangChain agents and connecting them to Telegram. Useful as a template for a chat agent that users reach directly in the messenger.
Autonomous HR Chatbot
An example project for an agent that answers HR-related questions on its own, using the tools available to it. It illustrates a concrete use case for internal assistants.
LLama Cpp Agent
A framework intended to simplify working with large language models via llama.cpp. Suited to integrating locally run models into chat and agent applications.
Letta
A platform for stateful agents with an extended memory that are designed to learn and improve over time. Relevant when an agent needs to retain context over long periods.
joinly
A voice-first AI assistant for online meetings that takes an active part in the conversation and handles tasks while the meeting is still running. An example of agents that work by voice rather than text.
Gobii
An open-source platform for deploying and managing browser-use agents at scale. Control is via a conversational interface and an API.
ClaudeClaw
A Claude Code plugin that orchestrates persistent agents. Features include routing across Slack, WhatsApp and Telegram, OS-level sandbox isolation, structured memory and webhook triggers.
OpenAgent
A self-hostable personal AI assistant built on LLMs, RAG and agent loops. It supports computer and browser use, coding, many model providers, a visual workflow builder and MCP tools.
MateClaw
An open-source personal AI system built on Spring Boot and Spring AI Alibaba. A single application provides a web admin, desktop app, embeddable widget and several messaging channels that share the same agent.
Orkas
A local-first multi-agent desktop application in which a coordinating agent directs specialists for research, coding, data analysis, documents and media. MIT-licensed, with your own model keys.
Ouroboros
A self-hosted general-purpose agent with a durable identity and memory. It can modify itself subject to review, use teams of specialist subagents and run on the desktop or headless.
5dive
Lets you run a team of named AI agents on your own server, organised on an org chart with a shared backlog. The agents hand work to one another, while a person makes the decisions from a phone.
Raven
A multi-agent environment designed as a persistent, self-evolving system for collaboration across domains. Its focus is on recursive self-improvement of the agents.
Markus
An open-source platform for building and running teams of AI agents. It offers role-based agents, task delegation, persistent memory and communication between agents.
From tool to outcome
In AI-first coaching you clarify which outcomes AI agents should be accountable for in your go-to-market and within which framework. REVAGO then designs, builds and runs your system. You pay for results – not for activity.
Key concepts
Other categories
A neutral overview, not a recommendation of individual providers. REVAGO is not affiliated with the listed projects. Descriptions are based on the projects’ public information; features and licences may change. Selection based on awesome-agents (kyrolabs), As of 19 October 2026.