TMX: The open standard AI agent memory has been waiting for

The problem no one talks about: your agent's memories are prisoners.

If you build an AI agent today using Mem0, your memories are locked in Mem0. Switch to Zep? You lose everything. Move to a new framework? Start from zero.

This is exactly the problem email had in 1970. Every system had its own format. You couldn't send an email from one system to another.

Then SMTP was invented. And email became universal.

Today I'm publishing TMX v0.1 — the SMTP of AI agent memory.

What is TMX?

TMX (Truvem Memory eXchange) is an open, model-agnostic JSON format for storing, exporting, and importing AI agent memories across any platform, framework, or provider.

It looks like this:

{
  "tmx_version": "0.1",
  "exported_at": "2026-06-26T20:00:00Z",
  "source": "truvem",
  "agent_id": "my-agent",
  "memories": [
    {
      "id": "550e8400-e29b-41d4-a716-446655440000",
      "content": "User prefers dark mode and concise responses",
      "created_at": "2026-06-01T08:30:00Z",
      "updated_at": "2026-06-01T08:30:00Z",
      "expires_at": null,
      "tags": ["preference", "ui"],
      "source_model": "gpt-4o",
      "metadata": {}
    }
  ]
}

That's it. Plain JSON. Human-readable. Portable.

Why this matters

Right now, the AI agent ecosystem is exploding. Every week there's a new memory provider, a new framework, a new cloud service.

But every one of them uses a proprietary format.

This means:

• Developers are locked to their first choice forever
• Agent memories can't travel between clouds
• Switching providers = losing everything your agent learned

This is the biggest hidden tax in the agentic AI stack.

TMX fixes it with a single open spec that anyone can implement — for free, with no approval needed.

The 5 core principles

1. Open — No license required. Implement TMX in any product, commercial or otherwise.

2. Model-agnostic — Works with GPT-4, Claude, Gemini, Mistral, Llama, or any future model.

3. Framework-agnostic — LangChain, CrewAI, Mastra, AutoGen — doesn't matter.

4. Human-readable — Plain JSON. No binary formats. Inspectable by any tool.

5. Minimal — The entire spec fits in one page. Complexity is optional.

Using TMX with Truvem today

Truvem is the reference implementation of TMX. You can start using it right now:

1. Get your free API key

Sign in with GitHub or Google at truvem.github.io/truvem/login.html

2. Install the SDK

pip install truvem

3. Write and read memories

from truvem import Truvem

client = Truvem(api_key="your_key")

# Write a memory
client.remember(
    agent_id="my-agent",
    content="User prefers dark mode"
)

# Read memories — from any session, any model
memories = client.recall("my-agent")
print(memories)
# {"status": "ok", "memories": [...]}

4. Export as TMX

curl -X GET https://truvem.onrender.com/v1/memory/export?agent_id=my-agent \
  -H "x-api-key: your_key"

You get back a valid .tmx.json file you can import anywhere.

Who should implement TMX?

If you're building:

• A memory provider (like Mem0, Zep, Letta)
• An agent framework (like LangChain, CrewAI, Mastra)
• A cloud AI service (like AWS Bedrock, Google Vertex)

You should implement TMX. Your users will thank you.

The requirements to be TMX-compatible are simple:

1. Export memories in valid TMX JSON format
2. Import valid TMX JSON files without data loss
3. Preserve all required fields

No fee. No approval. No committee. Just implement it.

The roadmap

The big picture

In 10 years, there will be more AI agents running than humans on Earth. Every one of them will need persistent memory.

The question is: who controls that memory?

If every provider keeps a proprietary format, agent memories become the new walled gardens — like email before SMTP, like the web before HTTP.

TMX is a bet that open standards win. They always have.

Read the spec

The full TMX v0.1 specification is published on GitHub:

👉 github.com/truvem/truvem/blob/main/TMX.md

Try Truvem (the reference implementation):

👉 truvem.github.io/truvem

What do you think? Is an open standard for agent memory something the ecosystem needs? I'd love feedback on the spec — open a GitHub issue or reply below.

Built by Dieng Amine — solo founder, Abidjan, Côte d'Ivoire.
Twitter: @gettruvem