API Live — Free during beta

The proof layer for
every AI action.

Every action your AI takes — logged, authorized, verifiable. 2 lines of code. Any model.

Get free API key View API docs
2
Lines of code
5+
LLMs supported
<100ms
Response time
$0
During beta
SHA-256
Append-only proof

Log every AI action with immutable proof.

Same result as before — plus a verifiable, append-only audit trail. Every action recorded, every authorization tracked.

API live · truvem.onrender.com
truvem_actions.py
# Before Truvem
result = openai.chat(messages)

# After Truvem — same result + immutable proof
result = truvem.log_action(
  agent_id="my-agent",
  model="gpt-4o",
  authorized_by="user@company.com",
  prompt=messages,
  result=result
)
# Returns: action_id + proof_hash
# (SHA-256, append-only)
Why it matters

Every action, accounted for.

Three pillars of trust for every AI action your agents take.

Recorded
Every action logged with a precise timestamp. Full prompt, full response, full context — nothing lost, nothing hidden.
Authorized
Who approved it, what scope, what permissions. Every action is tied to an identity and a clear authorization chain.
Provable
SHA-256 proof hash, append-only, forever. Cryptographically verifiable. Tamper-evident. Audit-ready from day one.
Why Truvem

Without memory vs with Truvem.

See the difference persistent memory makes for your AI agents.

Scenario ❌ Without Truvem ✅ With Truvem
User returns after 1 week Agent forgets everything. Starts from zero. Agent remembers preferences, history, context.
Switch LLM model All memory is lost. Locked to one model. Memory is model-agnostic. Works with any LLM.
Multi-device usage Context is fragmented across sessions. One memory layer for all devices and sessions.
User experience Repetitive. Frustrating. Feels dumb. Continuous. Personal. Feels intelligent.
Developer setup Build custom memory from scratch. Weeks of work. 2 API calls. Working in 5 minutes.
Use cases

Built for every AI agent.

From customer support to personal assistants — Truvem gives any agent persistent memory.

🎧
Customer Support Agent
Remember past tickets, user preferences, and conversation history. Never ask the same question twice.
Support
💼
Sales Agent
Track prospect interactions, objections handled, and deal context across every touchpoint.
Sales
🧠
Personal AI Assistant
Remember user habits, preferences, goals and context. Build an assistant that truly knows you.
Personal
📊
AI CRM
Store customer profiles, interaction history, and relationship context for every contact automatically.
CRM
💬
Chatbot with Memory
Give your chatbot persistent context so users never have to repeat themselves across sessions.
Chatbot
🤖
Autonomous Agent
Give long-running agents a persistent state layer. Remember tasks, decisions, and learned patterns.
Agents

Memory in 3 lines of Python.

Register once. Write memories. Read them back — from any session, any model, any machine.

API live · truvem.onrender.com
truvem_demo.py
from truvem import Truvem

# One-time setup
client = Truvem(api_key="your_key")

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

# Recall — any session, any model
memories = client.recall("my-agent")
# {"status":"ok","memories":[...]}
How it works

Simple by design.

No config. No embeddings. No infrastructure to manage. Just memory.

1

Sign in once

Sign in with GitHub or Google. Get a verified API key instantly. No credit card.

2

Write a memory

Call /memory/write with an agent ID and any string content.

3

Recall anywhere

Call /memory/read from any session, model, or device. Forever.

Compatibility

Works with every major model.

Model-agnostic by design. Pure REST — works with any language or framework.

GPT-4
OpenAI
Claude
Anthropic
Gemini
Google
Mistral
Mistral AI
Llama
Meta
Any model
REST API
Open Standard

Introducing TMX Protocol.

Agent memories should be portable. Like email with SMTP, Truvem is built on TMX — an open standard anyone can implement.

TMX v0.1 — Published June 2026
TMX: Truvem Memory eXchange
An open JSON format for AI agent memory portability. Store memories in Truvem, export to any system. Like HTTP for the web, but for agent memory. Free to implement. No approval needed.
Read the spec Read the article
Pricing

Simple. Transparent.

Beta — Free forever
$0
per month, free during beta
Unlimited memory writes & reads
Works with any AI model or framework
REST API + Python SDK
No credit card required
TMX export/import included
Get your free API key →
For developers

Add the badge to your project.

Show that your agent is powered by Truvem. Join the ecosystem.

Powered by Truvem ← Your badge will look like this
<a href="https://truvem.github.io/truvem">
  <img src="https://raw.githubusercontent.com/truvem/truvem/main/badge-1.svg" alt="Powered by Truvem">
</a>
Get started

Get your free API key.

Sign in with a verified account. Takes 10 seconds. No credit card.

Continue with GitHub
or
Continue with Google

1 verified account = 1 free API key. No abuse possible.

Then: pip install truvem