# Model-Agnostic AI Agent Tools

> Source: https://theagentsindex.com/collections/model-agnostic-ai-agent-tools (curated, quality-gated, re-verified)

A model-agnostic AI agent is one whose own listing names multiple model providers or an explicit bring-your-own-key (BYOK) layer, rather than baking in a single vendor's models. The set below is the buyers' shortlist when the question is "what runs everywhere" — agents whose own product documentation names Claude, GPT, Gemini and at least one open-weight family, and where swapping the underlying model is a configuration change rather than a rewrite. Per-entry notes name which providers each agent explicitly supports, so the gap between "model-agnostic in name" and "model-agnostic in the listing's own text" is visible at a glance.

| Name | What it is | Why it's here | Pricing | API |
| --- | --- | --- | --- | --- |
| [Agno](https://theagentsindex.com/agno.md) | An Apache 2.0 Python agent SDK and AgentOS runtime, with a Control Plane you run in your own cloud. | One of 20+ providers reachable through its model-agnostic Python interface, alongside OpenAI, Ollama and others. | — | Yes |
| [Aider](https://theagentsindex.com/aider.md) | Open-source AI pair programming in your terminal, free under Apache 2.0, built to run against 14 LLM providers | Anthropic is a named provider in its bring-your-own-key setup, alongside OpenAI, DeepSeek and local models. | — | No |
| [Botpress](https://theagentsindex.com/botpress.md) | An open-source (MIT) AI agent platform for customer support, with bi-directional MCP support, webchat, WhatsApp and voice channels. Self-host free or use the managed cloud from $0/month. | Anthropic Claude is one of the model-agnostic options, alongside OpenAI, Google Gemini and open-weight models. | — | Yes |
| [Cline](https://theagentsindex.com/cline.md) | The open-source coding agent. Bring your own key. Runs in your editor, terminal, and as an SDK, with optional ClinePass for flat-rate open-model access. | Anthropic is a named provider in its bring-your-own-key setup, alongside OpenAI or a local Ollama model. | — | Yes |
| [Dify](https://theagentsindex.com/dify.md) | Open-source, visual platform for building production LLM apps and agents, self-host it or run it in the cloud. | One of hundreds of model-agnostic options it can connect, alongside OpenAI, Google and open/local models. | — | Yes |
| [Google Agent Development Kit](https://theagentsindex.com/google-adk.md) | Google's open-source, code-first agent SDK with five language runtimes (Python, TypeScript, Go, Java, Kotlin), native A2A and MCP support, and an optional managed runtime on Google Cloud. | Reachable via LiteLLM alongside 100+ other providers, on top of native first-party Gemini support. | — | Yes |
| [LangChain](https://theagentsindex.com/langchain.md) | The most widely-used framework for building LLM apps and agents, with LangGraph for stateful multi-agent control: open-source and free to self-host. | The most widely-used multi-provider LLM framework, with integrations for Anthropic, OpenAI, Google, Mistral, Ollama and dozens more. | — | Yes |
| [LangGraph](https://theagentsindex.com/langgraph.md) | Open-source, MIT-licensed framework for building stateful, multi-agent AI workflows on cyclical graphs. | Anthropic is a named provider in the standard model interface its agent runtime exposes, alongside OpenAI, Google and others. | — | Yes |
| [Letta](https://theagentsindex.com/letta.md) | Letta is the open-source framework behind Letta Code, a stateful coding agent whose memory, skills, and identity persist across sessions rather than rebuilding context on every call. | Free tier runs stateful agents with bring-your-own-keys for OpenAI, Anthropic, Google and any OpenAI-compatible local endpoint. | — | Yes |
| [LlamaIndex](https://theagentsindex.com/llamaindex.md) | Open-source data framework for LLM apps, with LlamaParse as the paid agentic OCR for messy documents and LlamaAgents for document workflows. | Multi-provider LLM framework with Anthropic, OpenAI, Google, Mistral, DeepSeek, Ollama and Bedrock integrations wired through its agent stack. | — | Yes |
| [Microsoft Agent Framework](https://theagentsindex.com/microsoft-agent-framework.md) | Microsoft's unified, MIT-licensed agent SDK for .NET and Python: the supported successor to AutoGen and Semantic Kernel. | A named provider alongside Azure OpenAI, OpenAI, Ollama and Gemini (via its Go SDK). | — | Yes |
| [Microsoft AutoGen](https://theagentsindex.com/microsoft-autogen.md) | Microsoft's open-source framework for orchestrating conversations between multiple AI agents. Now in maintenance mode. | Free, MIT-licensed multi-agent framework where you pay only your chosen model provider — OpenAI, Azure OpenAI, Anthropic, Ollama and others. | — | — |
| [MindStudio](https://theagentsindex.com/mindstudio.md) | No-code visual builder for AI agents, with 200+ models, 1,000+ integrations, and one-click deployment as APIs, webhooks, or MCP servers. | No-code visual builder for AI agents, with 200+ models from Anthropic, OpenAI, Google, Mistral, Meta and open-weight providers. | — | Yes |
| [n8n](https://theagentsindex.com/n8n.md) | Source-available workflow automation with native AI-agent nodes and 500+ integrations. Self-hostable free, or on n8n Cloud from €20/month. | A named model-agnostic option alongside OpenAI and Ollama that its AI nodes can call. | — | Yes |
| [OpenAI Agents SDK](https://theagentsindex.com/openai-agents-sdk.md) | OpenAI's lightweight open-source framework for building agents in Python and TypeScript, with handoffs, guardrails, tracing and a sandboxed workspace mode for long-running work. | Free and MIT-licensed framework — you pay only for the model API calls your agents make, to OpenAI or any other provider. | — | Yes |
| [OpenCode](https://theagentsindex.com/opencode.md) | A free, open-source (MIT) AI coding agent for the terminal, desktop and IDE, with 75+ model providers, built-in subscription login, and a tested model marketplace behind it. | Reachable through its 75+-provider Models.dev integration, alongside GPT, Gemini and local Ollama models. | — | Yes |
| [OpenHands](https://theagentsindex.com/openhands.md) | Open source platform for cloud coding agents. Free local OSS, free Individual Cloud tier, self-hosted Enterprise with SSO and RBAC. | One of the two providers its own text names by example — Claude or GPT — for a bring-your-own-key setup that's otherwise model-agnostic. | — | Yes |
| [Pydantic AI](https://theagentsindex.com/pydantic-ai.md) | The Python AI SDK for type-safe agents, realtime voice, image generation and embeddings, with every model a string swap away. | A named provider in its type-safe, model-agnostic interface, alongside OpenAI, Gemini, Mistral and Ollama. | — | — |
| [Vapi](https://theagentsindex.com/vapi.md) | The developer platform for voice AI agents, bring your own model, voice and telephony, and Vapi orchestrates the call. | A named provider — with OpenAI and Google — in its otherwise bring-your-own-model pipeline. | — | Yes |
| [Zed](https://theagentsindex.com/zed.md) | The open-source, GPU-native code editor built from scratch by Atom's creators — runs Claude Code, Codex and other agents via the Agent Client Protocol it created. | First-class, named support alongside OpenAI, Google, Mistral, DeepSeek and more — and it natively runs Claude Code as an external agent via the ACP it created. | — | — |

## How this list is built

A tool earns a place here only when its own Published, sourced listing sets `attributes.modelAgnostic` to Yes — meaning the tool's own documentation names multiple model providers (Anthropic, OpenAI, Google, Mistral, Meta, DeepSeek and so on) or exposes an explicit bring-your-own-key layer where any compatible API endpoint can be plugged in. A single named provider with a roadmap promise of more does not qualify. Neither does a vendor's general "we work with all models" marketing claim that does not surface in the listing's own fields. Membership is derived from each listing's already-researched, gate-passed `attributes` and `integrations` data, not from inference about a tool's flexibility. Where a tool's claim to multi-provider support lives only in a third-party blog post or a community report, that is not enough; the listing must surface it in a field the listings lane has already verified.