Compare: Memory for one app, or knowledge every agent can trust

ContextNest vs Letta

Letta gives an agent its own memory that it improves as it works. ContextNest gives every agent the same knowledge a person has approved.

vsLetta

LettaWhat it is

Letta

An agent platform from the creators of MemGPT. Each agent keeps its memory as files in a git repository and updates them as it learns.

Pricing: Free, Pro $20 per month, API Plan $20 per month plus usage, Teams Pro $20 per seat per month. Enterprise is custom, with role-based access control and SSO (checked November 2026).

ContextNest is the company's approved knowledge, owned by people and read by every agent.

Sources

Choose Letta when you want a long-running agent that learns from its own work and manages its own memory.

Where Letta wins

Agents that learn between sessions
An agent saves what it learns as it works, and background "dreaming" runs fold recent conversations into memory without interrupting it.
Memory you can open and read
Memory is Markdown in a git repository, so each change is a commit you can inspect.
Your choice of model
Anthropic, OpenAI, Gemini, Bedrock and local models through Ollama all work, and you can switch mid-conversation.

Where Letta strains

Each one shows up when the same knowledge has to serve many people, many agents and more than one model.

Agents review their own memory
Letta's optional review step has the agent check its own updates, and its docs say this does not ask you for approval.
Other tools get replies, not memory
Claude Code or Cursor can message a Letta agent over MCP, but the six hosted MCP tools do not return what's in its memory.
Shared memory needs Letta Cloud
Several agents can share a repository, but Letta hosts it and only cloud-hosted agents can use it.

Side by side

ContextNest and Letta, on the questions that decide it.

The question	Letta	ContextNest
Every agent reads itCan Claude, ChatGPT, Cursor and your own agents all read the same knowledge?	PartlyRuns on most major models. Other tools can message a Letta agent but can't read its memory.	YesClaude, Cursor, PromptOwl agents and any MCP client read the same knowledge.
A person approves firstDoes a named owner approve knowledge before an AI may use it?	NoAgents write their own memory. The optional review is done by the agent.	YesTurn on stewardship and named owners approve every change before agents can use it.
Who sees what, enforcedIs access decided before the AI retrieves anything?	PartlyAn agent can't reach another agent's memory by default. Role-based access control is on Enterprise.	YesAccess keys are scoped to a nest, and restricted documents reach only the people named on them. Give each agent a nest that matches its job.
Record of reads and versionsCan you see who read what, and rebuild what it said on a past date?	PartlyEvery memory edit is a git commit. The docs describe no record of who read what.	YesEvery read is recorded: which person or API key, which document, and when. Every approved version is kept.
Survives a switchDoes the knowledge stay when you change models, vendors or people?	YesMemory is Markdown in git, and you can change an agent's model at any time.	YesPlain-text export, always. Your knowledge stays readable without us.
Self-hosted optionCan it run on infrastructure you control?	YesRun agents locally or on a self-hosted App Server. Shared memory needs Letta Cloud.	YesThe governed knowledge layer runs on your infrastructure, or we host it.

Scroll the table sideways. The first column stays put.

Get a second opinion

Each link opens the assistant with a neutral comparison question filled in. The answer comes from the assistant, not from us.

Ask ChatGPT Ask Claude Ask Perplexity

Last reviewed November 2026. Based on each vendor's public website and documentation, and prices show the month we checked them. Products change quickly, so check the vendor's current documentation before you decide. See something out of date? Tell us and we'll fix it.Product names and logos are trademarks of their owners; used for identification only; no endorsement or affiliation implied.

Use both

How they work together

Let each Letta agent keep what it learns from its own work. Connect it to ContextNest over MCP for the policies, pricing and product facts a person has approved.