Skip to content
Agent Search Engine.
Submit a tool
Menu

Alternatives · Agent Frameworks

Letta (formerly MemGPT) alternatives

These records overlap with Letta (formerly MemGPT)'s recorded tasks and have the same product type. The selection uses task text, then information completeness; stars and sponsorship do not set the order. Check the specific capability you need before switching.

8 task-related records from 57 category candidates · all agent frameworks · how we rank

The product you're replacing

Letta (formerly MemGPT)FrameworkAgent Frameworks

Stateful agents and persistent memory tooling with Cloud, CLI and SDK interfaces.

Costs not checked · 25k stars · Open source · repository push Oct 2026

Visit Letta (formerly MemGPT) →Full Letta (formerly MemGPT) record

Letta (formerly MemGPT)'s record says to avoid it if: You require the retired native Letta Code Desktop app or assume the current Desktop PWA works without a Cloud account.

Building one yourself? How to choose an agent framework →

Task-related research candidates

agent-frameworkFrameworkA framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and.NET.Costs not checkedOpen sourcePython · recorded license: MITRecord checked · Recent push 14kGitHub stars

Shared terms in the records: building, tooling. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; agent-framework — Production multi-agent workflows in .NET or Python with Microsoft tooling.

OpenAI Agents SDK (TypeScript)FrameworkOpenAI's lightweight, provider-agnostic TypeScript framework for multi-agent workflows and voice agents.Costs not checkedOpen sourceTypeScript · recorded license: MITRecord checked · Recent push 4.0kGitHub stars

Shared terms in the records: building, sdk. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; OpenAI Agents SDK (TypeScript) — TypeScript teams building agents or voice agents on OpenAI's primitives.

mcp-agentFrameworkA composable Python framework for building agents from MCP servers and simple workflow patterns.Costs not checkedOpen sourcePython · recorded license: Apache-2.0Record checked 8.6kGitHub stars

Shared terms in the records: building, developer. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; mcp-agent — Python developers who want agents whose tools are MCP servers and whose control flow stays simple.

Agent Development Kit for Go (Google ADK)FrameworkGoogle's open-source, code-first Go toolkit for building, evaluating and deploying AI agents.Costs not checkedOpen sourceGo · recorded license: Apache-2.0Record checked · Recent push 8.9kGitHub stars

Shared terms in the records: building, cloud. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; Agent Development Kit for Go (Google ADK) — Go teams building agent services that will run on Google Cloud or elsewhere.

Agent Development Kit (Google ADK)FrameworkGoogle's open-source, code-first Python framework for building, evaluating and deploying AI agents.Costs not checkedOpen sourcePython · recorded license: Apache-2.0Record checked · Recent push 22kGitHub stars

Shared terms in the records: building, cloud. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; Agent Development Kit (Google ADK) — Teams building multi-agent systems, especially on Gemini and Google Cloud.

LangGraphFrameworkBuild stateful workflows with code-defined and model-driven steps.Framework code; models, persistence, tools and hosting cost separatelyOpen sourcePython · recorded license: MITRecord checked · Recent push 44kGitHub stars

Shared terms in the records: developer, stateful. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; LangGraph — Developers who need explicit workflow state, persistence and human intervention controls.

MetaHarnessFrameworkGenerate a branded agent harness with tools, scoped memory and release policy.MIT source; selected models, services and execution infrastructure cost separately.Open sourcerecorded license: MITRecord checked

Shared terms in the records: developer, memory. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; MetaHarness — Developers comparing reusable scaffolding for project-specific agent harnesses.

SuperAGIFramework<> SuperAGI - A dev-first open source autonomous AI agent framework. Enabling developers to build, manage & run useful autonomous agents quickly and reliably.Costs not checkedOpen sourcePython · recorded license: MITRecord checked 18kGitHub stars

Shared terms in the records: developer, tooling. These text matches do not establish equivalent capabilities. The recorded use cases differ: Letta (formerly MemGPT) — Developers building stateful assistants with inspectable memory across sessions and configured interfaces; SuperAGI — Provisioning and running concurrent autonomous agents with extensible tooling.

Browse all agent frameworks →

Frequently asked

How were alternatives to Letta (formerly MemGPT) selected?
Recorded task topics and meaningful shared terms qualify same-type candidates in the same category. Results are ordered by that overlap, then information completeness and a stable identifier. Sponsorship and GitHub stars do not participate. The method can miss suitable tools whose records lack task information.
Does inclusion mean a tested replacement?
No. These are catalogue-based research candidates. Check current sources, costs, permissions and deployment requirements, then test your own task. ASE has not run a performance benchmark of this shortlist.