InherenceTurn any log into a proof
Use Inherence’s open-source, zero-knowledge proof — it’s on GitHub, download and run it yourself.
Logs are prone to disclosing sensitive information. A proof replaces that exposure, protecting privacy while maintaining trust. One small artifact, checkable by anyone, that reveals whether the rules held and nothing else.
Step 1
Add a log
Logs live in Datadog, Grafana, Axiom, an AI-agent tracer…? Connect a source →
Pull straight from your platform with the open-source CLI — your credentials and your logs stay on your machine, nothing reaches Inherence. First install it once, then pick a source for the exact command:
# one-time — needs the RISC Zero toolchain (see the repo README) git clone https://github.com/Inherencelabs/logproof && cd logproof cargo install --path crates/logproof-cli # installs the `logproof` command
Hypothetical scenario
Your company, Meridian Capital, is running a fleet of AI trading agents.
- Orders placed — venue, amount, account, time
- Risk-desk approvals
- Price quotes by venue
- Database reads & writes, by table
Step 2
Confirm what you want to prove (vs keep private)
Add a log first.
Step 3
Building your Proof
This takes time because it runs on a general-purpose zkVM. Inherence’s proprietary method proves the same thing at millisecond scale, fast enough to run inline. See the faster version ↓
Step 4
Your proof is ready
What if the wrong move
was never possible?
Inherence is proof of conduct at the moment of action, enforcing your rules before each step. Not a record you check later. Think of it like escrow. In escrow, the money physically cannot move until the conditions are met. You're not trusting a log that says it moved correctly, you're trusting that the wrong move was never possible. And because the enterprise product runs on Inherence's own proprietary zero-knowledge method — not the off-the-shelf prover behind this tool — it generates proofs 1,400–4,700× faster, quick enough to sit inline, at agentic speed.
