Open source · forensic evidence for AI agents

Trace a network connection back to the AI agent step that made it.

Mnestiq records what your agents saw, decided and did as signed, tamper-evident evidence. When something goes wrong, you can name the run, the step, the input that drove it and who approved it, and prove the record wasn't altered afterwards.

pip install mnestiq
v1.1.1 · Apache-2.0 · Python 3.10+ · Anthropic, OpenAI, Azure OpenAI and compatible APIs · PyPI
One log line, tracedexample
  1. Firewall10.0.4.17:53120 → 198.51.100.9:25 · 48 KB · no agent identity
  2. Mnestiqagent support-bot · run f48cccfa · step 6 · sandbox container:a91f02 · operator acme-helpdesk
  3. Causeuntrusted input at step 4 (web, via fetch_page) → model requested send_email → approved by policy:email-auto-approve
  4. Evidencesealed by checkpoint · signer ed25519:6bec5262… (pinned) · corroborated by vpc-flow.log, conn.log
mnestiq dashboard · evidence.jsonl · runs on your machine
Mnestiq dashboard: an agent run with findings and an expanded LLM call
A prompt injection, step by step: untrusted web content reaches the model at step 4, the model requests send_email, a policy auto-approves it, and a connection follows. Every record is sealed by a signed checkpoint.
The gap

Your logs see connections. Your traces see prompts. Nothing connects them.

A firewall alert: a host that runs AI agents sent 48 KB to an outside mail server on port 25. The responder now has three questions, and today's tools answer none of them.

Q1

Which agent did this?

Firewall and flow logs show an IP and a port. They don't know about agents, runs or steps.

Q2

Why did it do it?

Was the instruction from the user, or from a web page, a document or another agent? That has to be captured when it happens, not reconstructed later.

Q3

Can we prove it?

Ordinary logs can be edited by anyone with write access. Evidence for a post-mortem, an insurer or a regulator needs to show it wasn't.

How it works

Record, verify, investigate.

Every record above is sealed by a signed checkpoint.
When someone edits the record

Built for the person holding the incident.

integrity check · failed
Mnestiq dashboard: integrity check failed on an edited approval record
Someone edited the approval to blame a person. The edit even made a finding disappear, but the record no longer matches what was signed, and the seal shows as broken.
What the responder gets

The run, the step, the input and the approval.

The same evidence file answers the three questions: which agent, why, and whether the record can be shown to be unaltered. Anyone can check it with the open-source verifier, without Mnestiq.

Two products

Open evidence. Professional detection and investigation.

Mnestiq recorder

The evidence format, recorder, verifier and a local dashboard. Free and open source, so you can audit it yourself.

  • An open spec and JSON Schema, aligned with OpenTelemetry GenAI conventions
  • Python SDK for Anthropic, OpenAI, Azure OpenAI and OpenAI-compatible models
  • Signing keys held outside the agent, in Azure Key Vault or a dedicated signing service
  • Independent RFC 3161 timestamps and signed heartbeats
  • Crash-safe, size-capped, fails open by default
  • Dashboard runs on your machine only
Security model

What it proves, and what it doesn't.

Evidence is only useful if its limits are known. We state them plainly in the open spec and in a published threat model, where every attack considered is backed by an automated test.

Proves
Records weren't edited, deleted, inserted or reordered after they were written, given a pinned signing key.
Doesn't prove
That a compromised host told the truth when it wrote them. Corroborate with independent infrastructure logs. The Workbench does this.
Gap
On its own, an evidence file can be cut back to an earlier checkpoint without detection. Keep a copy of the latest checkpoint somewhere else and verify against it to close this. The Evidence Vault does this for you.
Keys
Signing keys can be held in Azure Key Vault or in a signing service under a separate account. A compromised agent cannot copy the key. With the signing service it also cannot have earlier evidence signed again, because the service signs each chain only forward. Key Vault signs whatever the agent's identity asks for, so there the Evidence Vault's copy is what shows a rewrite.

Verify, don't trust.

You should not have to take a security vendor's word for it. The recorder and its verifier are open source, every claim in the threat model is backed by an automated test, and any evidence file can be verified independently, without Mnestiq. Prompts and data stay on your own systems: only signed checkpoint fingerprints leave, and only to destinations you configure.

Independent review is welcome. The break-it challenge provides a signed, timestamped incident record: alter it without the verifier noticing, and tell us how.

Design partners

Running agents with real permissions in production? Let's talk.

We're looking for a small number of security and platform teams to shape the workbench around real incidents.

alexander@mnestiq.com