# Volance

> Volance is an explainable behavioral bot- and agent-detection product by Oops Games LLC. It returns a human-likeness score, classifies each session as human, agent, or bot, explains the signals behind the verdict, and leaves enforcement to your own server.

Volance is a scoring service, not a wall. One reviewed script on your pages collects on-site behavioral and client signals; your backend relays the evidence with a secret key; Volance returns a score, a classification, a verdict, and the evidence behind it. Your server decides whether to allow, flag, or block. There is no proxy or CDN in the request path.

HTML version: https://volance.com/llms

## Key facts

- Classification labels: `human`, `agent`, `bot` — distinct from the policy-facing `verdict`.
- Score: a human-likeness estimate on a 0–100 scale. An estimate, not a probability or proof of identity.
- Verdict bands (target contract): human at 80 or above, likely human 70–79, suspicious agent 40–69, bot below 40.
- Suggested action: allow, flag, or block — recommended by Volance, enforced by your server.
- Integration scopes: a whole site, selected pages, or a single server-side API endpoint.
- Authentication: a public `pk_` key in the browser requests a session manifest; a secret `sk_` key stays server-side, sent as `Authorization: Bearer`.
- Privacy: on-site signals only, no cross-site identifiers, hashed IP addresses, 30-day raw retention, opt-in collection, and no raw keystrokes or typed content.
- Framing: behavioral detection, not a CAPTCHA; a score does not prove a human was present.

## Documentation

- [Integration overview](https://volance.com/docs): how the collector, the API, and the portal fit together, the three scopes, and the request fields.
- [API reference](https://volance.com/docs/api): endpoints, request and response fields, verdict thresholds, signal layers, Web Bot Auth, errors, and rate limits.
- [Portal guide](https://volance.com/docs/portal): projects, keys, the consented test area, live traffic, sessions and evidence, usage and plans, roles, audits, and installing the collector.
- [Agent monetization](https://volance.com/docs/agent-monetization): the 402/x402 paygate — classify the caller, verify Web Bot Auth trust, set a price, and enforce the decision on your own edge. Volance does not proxy your traffic and never holds funds.

## Atlas

- [Global agent activity](https://volance.com/atlas): a public, aggregate view of agent and bot activity — the human/agent/bot split, bot categories and operators, and the x402 economy. Estimate, updated daily.
- [State of agent traffic](https://volance.com/atlas/report): the dated monthly report — automated traffic, AI crawlers, top bots, and the x402 economy.
- [Atlas method](https://volance.com/atlas/method): sources, the k-anonymity gate, and the limits.
- [Atlas facilitators](https://volance.com/atlas/facilitators): x402 volume by facilitator, with organic and raw figures.
- [Atlas services](https://volance.com/atlas/services): the Coinbase CDP Bazaar catalog — services, listed prices, networks, and calls.
- [Atlas AI agents](https://volance.com/atlas/ai-bots): top AI user agents by request share, from Cloudflare Radar.
- [Atlas widget](https://volance.com/atlas/widget): an embeddable view of agent activity and the agent economy.

## Product

- [How it works](https://volance.com/how-it-works): collection, scoring, and the decision that stays on your server, from one endpoint to the whole site.
- [Agent monetization](https://volance.com/agent-monetization): get paid for agent traffic — classify the caller, verify Web Bot Auth trust, set a price, and return an x402 challenge your edge enforces. Volance never holds funds.
- [Pricing](https://volance.com/pricing): Free (10,000 customer scores a month), Pro ($99/month, 250,000), and Business plans set up with your team.

## Trust

- [Privacy](https://volance.com/privacy): what the website and the service collect, what is never collected, retention, sharing, and your controls.
- [Terms](https://volance.com/terms): terms for the website and the service, including plans, keys, and acceptable use.

## Company

- [About Volance](https://volance.com/about): what it is, who makes it, and where to find us.
- Volance is a product of Oops Games LLC. This website is informational; accounts and scoring happen in the developer portal.
- [Developer portal](https://app.volance.com): sign-up, API keys, consented tests, and live traffic.
- Contact: help@volance.com. X: @volancehq. GitHub: github.com/oops-games-llc.

## Machine-readable files

- [llms.txt](https://volance.com/llms.txt): the concise text index.
- [llms-full.txt](https://volance.com/llms-full.txt): the page-by-page text index.
- [HTML version](https://volance.com/llms): the same guide as a readable page.
- [sitemap.xml](https://volance.com/sitemap.xml): every indexable URL.
- [feed.xml](https://volance.com/feed.xml): the blog feed.
- [Atlas snapshot](https://volance.com/atlas/snapshot.json): the Atlas data snapshot.

## Comparisons

- [Volance vs Google reCAPTCHA](https://volance.com/compare/recaptcha): A graded, explainable score vs. a single opaque risk value.
- [Volance vs Cloudflare Turnstile](https://volance.com/compare/cloudflare-turnstile): Explainable scoring vs. a privacy-friendly pass/fail challenge.
- [Volance vs Fingerprint](https://volance.com/compare/fingerprint): On-site behavioral scoring vs. persistent device identification.
- [Volance vs Detection Lab](https://volance.com/compare/detection-lab): The closest match on approach — so this one is about scope and governance.

## Use cases

- [Whole-site visibility](https://volance.com/use-cases/whole-site): Point tools tend to cover one form. Volance gives you a consistent read on every session across a shared template, without proxying your traffic.
- [Checkout & signup](https://volance.com/use-cases/checkout-and-signup): Signup and checkout are where abuse costs you directly. Volance scores those sessions and hands your server the evidence before the account or the order is created.
- [Marketplaces](https://volance.com/use-cases/marketplaces): Marketplaces sit between genuine participants and a lot of automation. Volance helps you tell them apart with evidence, and keeps the enforcement decision yours.
- [API & trial abuse](https://volance.com/use-cases/api-and-trial-abuse): Not every request has a browser. For APIs and trials, Volance scores the evidence your edge already has and returns it to your server — no fabricated browser signals.
- [Publishers & content](https://volance.com/use-cases/publishers): Publishers need discovery and pay attention to scraping. Volance separates the crawlers you want from the ones you do not, with evidence behind each call.

## Blog

- [Checkout and signup: which behavioral signals actually help](https://volance.com/blog/checkout-and-signup-signals): Fake accounts, card testing, and coupon abuse hit the same two pages. Here is what behavioural signals can tell you.
- [API abuse without a browser: scoring request evidence](https://volance.com/blog/api-abuse-without-a-browser): APIs and free trials get farmed, but there is no page to instrument. Score the request evidence your edge already has.
- [How Web Bot Auth works, step by step](https://volance.com/blog/how-web-bot-auth-works): A practical walkthrough of signed agent requests: what is signed, where the key lives, and how a verifier decides.
- [We red-teamed our own detection. Here is what the dashboard showed.](https://volance.com/blog/red-teaming-our-own-detection): We pointed our red team at a live site we operate, then watched the classification, decisions, and notional toll land in the dashboard.
- [What share of your traffic is automation? The number is not the point](https://volance.com/blog/what-share-of-traffic-is-automation): Every major 2026 report puts automation at or above half of web traffic. The useful question is what kind of automation it is.
- [Measurement is not proof (and why that is fine)](https://volance.com/blog/measurement-is-not-proof): A detection score is an estimate, not a verdict of identity. That is not a weakness — it is the only honest way to describe what session scoring can do.
- [From a bot score to a policy: allow, flag, block](https://volance.com/blog/from-score-to-policy): A score is not a decision. This is how the bands, verdicts, and actions fit together, and why enforcement belongs on your server.
- [Signed vs spoofed agents: Web Bot Auth in practice](https://volance.com/blog/signed-vs-spoofed-agents): How verified, spoofed, and unresolved Web Bot Auth signers differ — and why an unknown signer should never be punished.