The oversight kit

Review AI proposals before they change your data.

With the oversight kit, a person reviews each model proposal before it changes any data. Applied changes are logged and can be reversed. You can verify the log to check for tampering. Before each model call, the system checks which data the model may receive. Everything runs in your own environment.

An agent prepares a proposal. A person reviews it, and their decision is recorded.

You install the libraries, released under the MIT licence from November 2026, and run them on your existing Postgres database. We do not host the kit or process your data. You can also run the model in your own environment.

Diese Seite auf Deutsch

How it works

From proposal to recorded decision

01
Check content
Content is classified as public, internal or confidential before each model call. Confidential content stays with a model in your environment. Internal content is redacted before it is sent outside. The record of each call keeps a digest of the content sent, rather than the content itself.
02
Propose
The model returns a typed proposal for existing records. It includes the proposed changes, the current values, the proposer and the person or organisation they are acting for. It does not write directly to the database.
03
Decide
A designated reviewer checks the proposal in your application. They can edit it, accept individual changes or reject it. If the underlying record has changed, the reviewer sees that before applying the proposal.
04
Record
A hash-chained ledger records each applied change, who made it and on whose behalf. Reversing a change adds a new entry and preserves the earlier records. The latest hash is regularly secured outside the database, so even a fully rewritten chain shows up at the next check.

Who it is for

Software vendors and regulated operators

Software vendors whose AI takes part in regulated decisions

If your product uses AI in recruitment, education, credit, insurance or access to services, your customers ask how people supervise its decisions and what the system records.

The obligations for high-risk systems listed in Annex III of the AI Act, among them automatic record-keeping (Art. 12) and human oversight (Art. 14), apply from 2 December 2027. That date comes from the 2026 amendment.

Human review and a tamper-evident change log integrated into your own product.

Regulated operators whose officers sign off in person

In critical infrastructure and in the financial sector, new software is assessed before it goes live. The security and data protection officers gather supplier evidence and each prepare an assessment before approving its use.

§ 30 BSIG counts supply-chain security and secure procurement among the minimum measures, and under § 38 management has to approve and oversee them. DORA requires a register of ICT third-party providers. The GDPR requires processing agreements, a record of processing activities and, where the risk is high, an impact assessment.

A model gathers the evidence and links each statement to its source. The responsible person reviews and signs the result. The kit runs in your environment; you can keep the model there too.

Your legal adviser can assess whether your system is high-risk and which obligations the kit can help you meet. We provide technical implementation, not legal advice.

What exists today

The packages

Each package has its own documentation explaining how to use it and where its limits are. We can build the application around these packages for your workflow.

@octabits-io/proposal
Defines typed proposals for existing records, including previous values and who proposed each change. Supports partial acceptance, checks for changes to the underlying data, application and reversal of proposals, and storage while a decision is pending.
@octabits-io/agent-ui
Manages review state independently of the UI framework, with bindings for Vue and React. The review screen is integrated into your application.
@octabits-io/agent-ledger
Records each agent action, on whose behalf it was taken and how to reverse it. Entries form an append-only hash chain in Postgres. Each storage implementation is checked by the same conformance tests.
@octabits-io/disclosure
Checks content before model calls. A lint rule fails the build if a model call bypasses that check.
@octabits-io/server
Provides the review process over MCP. Tools that change data must return a proposal unless the host explicitly permits direct writes. Each call is logged with the caller’s identity.
octaflow
Runs durable multi-step workflows on Postgres. The model integration defaults to an EU-based provider and also supports self-hosted models.

The engagement

Implementation in your system

In your existing system

One AI workflow in your system: reviewable, logged and reversible.

We implement one workflow in which a model prepares a proposal for human review, such as a vendor assessment, a record update or a document awaiting signature. Using the oversight kit, we add a review screen, a hash-chained change log with revert support, and checks on the content sent to each model call. The workflow runs in your environment on your Postgres database. You own the application code.

You get: A working workflow in your environment, tests and documentation of which data the model may access. We measure the time per case and how much reviewers change each proposal to help you assess the results.
Duration: 8–10 weeks
Pricing: Fixed price, agreed after a first call; optional ongoing support

You own the application code and run it in your environment. We implement the workflow, maintain the libraries and can provide ongoing support.

Not yet

What the kit does not do yet

Automatic generation of documentation from the ledger and disclosure records. We currently prepare this documentation by hand as part of the implementation.
Storage of the full content of model calls for a defined retention period. The kit currently stores only a digest. Requirements to retain the full content need to be assessed for each installation; this feature is not yet implemented.
A ready-to-use application. We build the workflow for each customer using the kit, in their own codebase.

Discuss your workflow with us

Tell us what the model should prepare, who reviews the result and how the process works today. We can then discuss which workflow to implement first.

mail@dk2.eu Contact