Documentation

The Novera API

A workspace API key lets a script, a CI pipeline or an AI assistant read your workspace's agents, suites, runs and verdicts without signing in. A key can also start runs only if you tick Can also start runs when you create it — each run it starts is counted exactly like one started from the button: one of the trial's three runs, then one of your plan's or grant's runs, graded on your own model key if the workspace has one. No key can change a policy, approve anything, or publish or revoke a report.

The Ask Novera assistant

Every signed-in page has an Ask Novera button in the bottom-right corner. It answers questions about Novera and about your own workspace: your agents, your recent runs and their counts, and how many runs your trial has left.

Reports in CI, and verifying a copy

A sealed report can be read by a machine as well as a person. Every report link has four downloads — Markdown, CSV, JSON and JUnit XML — and the JSON and JUnit files carry an exit code a CI pipeline can act on. Nothing here can change a report, start a run or publish anything.

Connecting your agent, step by step (no developer needed for the first steps)

This page takes you from nothing to a first test of your AI support agent. Most of it needs no technical knowledge. Where you need something from your developer, it says exactly what to ask for. Developers will find the full technical detail in Connecting an agent.

Connecting an agent

Novera talks to your agent over HTTP. You give it the endpoint URL, a JSON request body template containing {{input}} where the scenario text should go, and the path to the reply text in the response.

Data and privacy

The database runs in Frankfurt, in the EU. Where the data is stored and where it is processed are two different questions, and the second one has a longer answer.

Why the report can be trusted

Results, policy versions, probe receipts and reports are append-only in the database itself, enforced by triggers rather than by application code. A failing case cannot be edited into a passing one or deleted from a run — not by you, not by Novera, and not with the administrative key. The one change allowed is the retention pass: after your retention period it empties an agent's raw reply (and a probe receipt's reply after 90 days), keeping the verdict and a SHA-256 of what was removed.

Finding your way around Novera

This page walks through Novera's screens once you have signed in: what each place is for, what you do there, and where to find the report you hand to a client. For what each part does and why, see Novera in plain words.

How a run works

A run names three things at the moment it starts: the agent, the policy version, and the suite version. It executes against those, not against whatever is current later.

Integration with GitHub Actions

This guide adds a workflow that tests your agent on every pull request and every push to main: it starts a Novera run for the commit, drives it to the end, and fails the job when the evidence fails or is incomplete. Set as a required status check, the job blocks the merge in the same cases. It uses only the public API described in the API; the exit codes are the ones in Reports in CI, and verifying a copy.

Integration with LangGraph (Python)

This guide wraps a LangGraph graph in a FastAPI endpoint that Novera can call, protected by a test credential in a request header, and returns the graph's tool calls so they can be graded. The Novera side is the HTTP connection described in Connecting an agent. The Python side was written against fastapi, langgraph and langchain-core as of 2026; the graph itself is your own, so only the parts that touch messages and tool calls are shown.

Integration with the Vercel AI SDK

This guide wraps an agent built with the Vercel AI SDK in a Next.js route handler that Novera can call, and returns its tool calls so they can be graded. Everything on the Novera side is the HTTP connection described in Connecting an agent: an address, a JSON body template, a path to the reply, and optionally a path to tool activity.

What Novera is not

Novera is evidence of testing. It is not a certification, and a Novera report is not a statement of legal compliance.

Novera in plain words: what each part does

Novera tests an AI support agent the way a careful auditor would: it sends the agent a fixed set of realistic customer conversations, checks every answer against the rules the company wrote for it, and seals the result in a dated report that anyone can check was not changed afterwards. This page explains each part in plain words, and why you would use it. For where each part sits on screen, see Finding your way around Novera. For the words, see the plain-words glossary.

Plain-words glossary

Every word you are likely to meet in Novera or in a Novera report, explained without jargon. The same words carry a small ? on screen that explains them in place. For the bigger picture, see Novera in plain words and Reading a report.

Reading a report (for clients and managers)

This page is for the person who receives a Novera report: a manager, a support director, a legal adviser, a client. You do not need to know how AI models work to read one. Each section below follows the report from top to bottom and says what it means in plain words.

Regressions from production

When your agent gets something wrong with a real customer, that conversation is the best test case you have. The Regressions page (under Scenarios) turns it into a scenario, and every later run shows whether the same mistake came back.

Scheduled runs

A schedule runs a suite against an agent again on a calendar, so a change in the agent — a new model version, an edited system prompt, a changed knowledge base — shows up in a comparison without anyone having to remember to press Run the suite. Set one up on the agent's page, under Schedule.

Signing in and your account

You sign in to Novera with an email address and a password, or — where they are switched on — with a Google or GitHub account. Every method opens the same account: the same workspaces, agents, reports and API keys.

The Suite Builder

The Suite Builder makes a suite you can defend without writing every scenario yourself. You start from a measured baseline, add your own documents, compare them with what your agent actually does, and approve what applies. Nothing a model drafts is tested until a person approves it, and nothing the builder drafts is presented as your policy.

The trial, plans, and using your own model key

Each account gets three suite runs free, to evaluate Novera and see a first sealed report. They are counted across every workspace you own, so a second workspace does not start a second trial. There is no card.

What Novera does

Novera runs a versioned scenario suite against an AI support agent you operate, grades every answer against a policy you wrote, and produces a dated report you can hand to a client.

What the report contains

The report is a single link, expiring and revocable, that prints cleanly to PDF.

Writing and importing a suite

A suite is a versioned list of scenarios. Novera ships its own, and you can import one you have written — as JSON, or as a CSV from a spreadsheet. Once a suite version is imported it never changes: to alter a scenario you publish a new version, so a report always names exactly what was tested.