FAQ
Questions about Anchor8.
What Anchor8 does, how a step is decided, and what happens to your data, in a sentence or three each.
Updated
Anchor8
What is Anchor8?
Anchor8 is a control layer for autonomous AI agents. It sits between an agent and its tools and checks every step both ways: the request on its way out and the result on its way back. Safe steps go through in milliseconds, dangerous ones are stopped, and a person is asked only when Anchor8 can't tell.
Who is Anchor8 for?
Anyone whose agents act on their own: one developer running personal agents, a small team, or a whole company putting agents into production. If your agents call tools, touch customer data, change production or move money, Anchor8 sits in front of those actions.
What is AI agent security?
AI agent security is keeping an autonomous agent from doing harm with the access it has been given. An agent that can call tools can send email, change records, wire money or delete a production database, and it can be steered by instructions hidden in what it reads. Securing it means checking each action before it runs, checking what the agent reads before it acts on it, and keeping a record of every decision.
How is Anchor8 different from guardrails?
Guardrails usually check the text a model reads or writes, often with rules, detectors or another language model. Anchor8 checks actions: every tool call and every tool result, with one model trained for that one decision and no language model in the path of a normal step. It learns from the answers you give, so it asks you less over time.
What is prompt injection, and does Anchor8 stop it?
Prompt injection is an instruction hidden in something an agent reads, such as a web page, an email or a file, meant to make the agent do what its user never asked for. Anchor8 checks tool results on their way back, so a result carrying instructions for the agent is stopped before the agent reads a word of it.
How a step is decided
What can happen to a step?
One of three things. It goes through, it is stopped, or it goes to a person. You only see the ones that need you.
What happens when Anchor8 stops a step?
The step doesn't run, and the agent is told why so it can find another way. A stopped step isn't the end of the run: the agent keeps working.
When does Anchor8 ask a person?
Only when a step needs judgement, and at no other time. The person's answer goes on the record, and the next time a step like that comes along, Anchor8 doesn't need to ask.
Does Anchor8 learn from my decisions?
Yes. Every answer you or someone on your team gives becomes training data for a model that is yours alone, never shared with another customer. Our base model learns from synthetic and public data only.
What does Anchor8 record?
An accountability record for every step: who acted, what it wanted, the data involved, what Anchor8 decided and why, and what happened next. Entries are chained, so a changed entry shows.
Does Anchor8 slow my agents down?
Safe steps go through in milliseconds. No language model sits in the path of a normal step, so the check stays fast enough to run on every one.
Does Anchor8 call my tools?
No. Your tools stay in your systems and Anchor8 never calls them. Your agent does, once a step is let through.
Your data
Where is my agents' data kept?
In storage on AWS, encrypted with your own key before it is written. Our logs never hold it. It is kept for as long as you're a customer, and when you leave, we delete your key, which makes every stored copy unreadable, and then delete the copies.
Security, said plainlyWho at Anchor8 can see my data?
Nobody by default. Access to one customer needs a written reason, lasts an hour and shows up in that customer's audit log. Sales and marketing never get it.
Is my data used to train anyone else's model?
No. Your review decisions train your own model and no one else's, and no Claude output goes into any training data.
Getting in
Can I use Anchor8 today?
Not yet. We're letting people in a few at a time. Take a spot on the waitlist and we'll write when it's your turn.
Join the waitlistWho makes Anchor8?
Anchor8 is built by Anchorate Private Limited, a company founded to make it safe to let AI agents act on their own.
Your agents are already running.
Anchor8 is for one developer, a small team or a whole company. We're letting people in a few at a time. Take a spot in the line and we'll write when it is your turn.

