QVeris
QVeris is a capability routing network that lets AI agents find, vet and call thousands of external APIs and data services through one protocol and a single key, with the cost of each call shown before it runs.
What is QVeris?
QVeris is a routing layer that sits between an AI agent and the thousands of external APIs and data services it might need. It is published by QVERIS AI, LLC. The problem it targets is familiar to anyone who has built an agent: without such a layer, every provider means another integration, another key, another quota and another contract.
The product reduces that to three stable verbs. Discover takes a plain-language query and returns ranked capability matches, free. Inspect returns the full schema, parameters, estimated latency, success rate and credit cost, also free. Call executes the capability in a sandbox and returns structured JSON, and it is the only step that consumes credits. Two further verbs sit alongside them: Probe validates parameters and returns a quote without spending anything, and an audit surface exposes usage history and a credits ledger.
The homepage claims 10,000+ capabilities across 15+ categories, 99.99% availability, sub-500 ms p95 latency and support for 27 agent platforms. These are vendor figures, with no public status page behind them. The catalogue leans heavily towards finance, with six highlighted domains: quant trading, macro and fixed income, risk and compliance (KYC, sanctions, beneficial-owner chains), investment research, crypto and digital assets, and alternative signals. The machine-readable llms.txt file announces wider coverage, naming search, weather, maps, documents, social, blockchain and healthcare.
Each capability is backed by several comparable providers, with quality signals, pricing rules and fallback routes, and a Provider Hub exposes latency and success rate per provider.
Access paths are ranked by the vendor: hosted MCP over Streamable HTTP for remote-capable clients; a CLI for agents with shell access, whose selling point is that tool schemas never enter the model's context; a local MCP server as a stdio fallback; TypeScript and Python SDKs; and a REST API with 26 documented operations. All client tooling is open source on GitHub and npm, while the routing engine itself stays a managed service.
What it does
- Search thousands of real-world capabilities in plain language, free of charge
- Inspect a capability's schema, latency, success rate and credit cost before committing to it
- Validate parameters and obtain a price quote with Probe, without spending a credit
- Execute a capability in a sandbox and receive structured JSON in return
- Audit every charge by execution ID through usage history and the credits ledger
- Compare the competing providers behind a capability on coverage, latency and billing rules
- Connect an existing agent in one command through the CLI, hosted MCP, an SDK or REST
When to use QVeris / When not to
A quick filter to help you decide if QVeris is the right fit.
When to use QVeris
- Developers building AI agents that need many third-party APIs without wiring up one integration per provider
- Owners of shell-capable agents, where the CLI keeps tool schemas out of the model context entirely
- Teams on MCP-compatible IDE clients who want a hosted endpoint rather than a local server process
- Finance and investment analysts covering quant signals, macro and rates, earnings research, crypto and alternative data
- Compliance, KYC and sanctions teams that need screening and beneficial-owner lookups on demand
When not to use QVeris
- Teams that need one well-known API only, where a routing layer adds cost without adding reach
- Regulated buyers who require a signed DPA, a named subprocessor list or a SOC 2 or ISO 27001 certificate
- Anyone looking for a mobile app, since there is no iOS or Android release
- Finance departments that need a predictable monthly invoice rather than variable credit consumption
- Non-technical users, because every access path runs through a CLI, an MCP client, an SDK or the REST API
How to use QVeris
A typical end-to-end flow, from setup to results.
- Create an account with an email and password, or through Google or GitHub; signing up credits 1,000 credits to your balance
- Generate an API key from Dashboard / API Keys and store it outside your source control
- Choose an access path: hosted MCP for remote-capable clients, the CLI for agents with shell access, a local MCP server, an SDK or the REST API
- For the CLI, run the one-line installer from the site or npm install -g @qverisai/cli, then authenticate with qveris login
- For hosted MCP, register the endpoint https://mcp.qveris.ai/mcp and send your key as an Authorization: Bearer header
- Run qveris doctor to confirm that Node.js, the API key, the endpoint and connectivity are all in order
- Find a capability with qveris discover "your intent" --json
- Review it with qveris inspect, or run qveris probe to validate parameters and get a quote without spending credits
- Execute with qveris call --params '{...}' --json; session shortcuts let you refer to earlier results by index for 30 minutes
- Confirm what the call actually cost by querying usage history with its execution ID once settlement has completed
Pros & Cons
Pros
- One API key and one protocol replace a separate integration for every provider
- The cost is knowable before you commit, since Inspect and Probe both return pricing without charging
- Discovery and inspection are free and unmetered, and credits never expire
- No subscription and no auto-renewal: you buy credits when you need them
- The CLI keeps tool schemas out of the model context, which matters in long agent loops
- Every charge is auditable by execution ID, with an explicit charge outcome rather than a guess
- All client tooling is open source and inspectable, and the documentation is written for machines as well as people
Cons
- No postal address is published anywhere on the site
- No GDPR mention, no DPA, no subprocessor list and no security certification such as SOC 2 or ISO 27001
- The headline figures are vendor claims, with no public status page or verifiable report behind them
- The site advertises 27 supported agent platforms but names only eight; the page that would list them all renders in JavaScript and returns nothing to a fetch
- A single call ranges from 1 to 100 credits, which makes spending hard to forecast before inspection
- A very young operation: the domain was registered in August 2025 and first archived in January 2026
- No mobile app, and the documented backend health endpoint points at localhost, so it is of no use to a visitor
Pricing & Plans
A permanent free plan is available: it grants 1,000 credits on signup plus 100 credits per daily login, with unused daily credits reset the following day, under a limit of 10 requests per minute. Discover, Inspect and Probe are free at every tier, and only Call consumes credits, at a rate of 1 to 100 credits per call. The lowest paid entry point is the Scale On-Demand top-up, which starts at USD 1.00, while the headline paid plan is Pro at USD 19.00 for 10,000 credits. Credits never expire and no monthly fee applies. Enterprise pricing is available on request only.
- USD 0 - 1
- 000 credits on signup plus 100 credits per daily login (unused daily credits reset the next day)
- 10 req/min
- basic tools access
- community support
- Live Demo access
- standard queue
- USD 19 - 10
- 000 credits
- 100 req/min
- USD 0.002 per credit overage
- access to all 10
- 000+ tools
- email support within 24 hours
- usage analytics
- from USD 1 - top up any amount
- everything in Pro
- volume bonus credits (+5% at USD 100
- +10% at USD 500
- +15% at USD 1
- 000)
- no monthly commitment
- credits never expire
- price on request - custom pricing
- dedicated support and custom SLAs through Contact Sales
Data, GDPR & hosting
A consolidated view of how QVeris handles your data.
GDPR overview
There is no mention of the GDPR anywhere on the site. Across the fifteen pages collected, neither that regulation nor any other is named, no legal basis is cited, no Article 27 representative is designated and no data protection officer is identified. What the privacy policy does offer, without tying it to any statute, is a set of rights in section 5: deletion of the account, deletion of the associated personal information, and access to or correction of it, all exercised by writing to contact@qveris.ai. Section 6 states that information may be transferred and processed outside the user's jurisdiction, including in the United States, under standard contractual clauses or equivalent safeguards. Section 8 restricts the service to users aged 13 and over. Both documents are dated 11 May 2026. No DPA is published or offered, and no subprocessor list exists.
Who owns the data?
Under the Terms of Use you keep ownership of everything you submit, which the document calls User Content: both the content and the queries themselves. Submitting it grants QVERIS AI, LLC a limited, non-exclusive, royalty-free licence, confined to operating and providing the service, and that licence explicitly covers passing query content on to third-party providers when a call requires it. You must warrant that you hold the rights to what you send. Ownership runs the other way for the platform itself: the software, APIs, trademarks and documentation belong to QVeris AI or its licensors, and the Terms grant no right to use its branding.
Reuse rights
The Terms confirm that you keep ownership of the content and the queries you submit, and they place no restriction on how you reuse the structured results a call returns. The licence you grant runs in the other direction only: QVeris AI may process and transmit your content solely to operate and provide the service, which includes forwarding query content to the third-party AI providers needed to fulfil a request. The privacy policy adds that usage logs and query content are not linked to individual user identities, that personal information is neither sold nor used for advertising, and that no cookies or similar technologies serve advertising or cross-site tracking. Training models on customer data is not mentioned anywhere in the published documents, in either direction.
Data retention & training
Hosting summary
Hosting information is thin. The only statement comes from section 6 of the privacy policy, which says information may be transferred to and processed in countries outside the user's jurisdiction, including the United States, relying on standard contractual clauses or equivalent legal safeguards. No specific hosting country, no region and no cloud provider is named anywhere in the published documents. The security page describes controls without identifying a host or a location: production services in private subnets with no direct public access to databases, AES-256 encryption at rest and TLS 1.3 in transit. One network observation, which the vendor does not declare and which should not be confused with its statement: the domain's IP address geolocates to Singapore, on an Alibaba Cloud autonomous system, with no anycast. That describes where the website is served from, not necessarily where customer data is processed. Anyone with a data residency requirement should have this confirmed in writing before committing.
Things to keep in mind
Risks and trade-offs to weigh before adopting QVeris.
- Query content is forwarded to third-party AI providers the site never names, so sensitive or regulated data should not be routed through it without checking first
- The absence of a DPA, a subprocessor list or any certification makes the tool hard to defend in a regulated procurement review
- The privacy policy points to processing in the United States while the site's own IP address resolves to Singapore on Alibaba Cloud, so where processing actually happens is not established
- An agent calling capabilities in a loop can burn credits quickly, since one call ranges from 1 to 100 credits: inspect before you call
- The vendor warns that a failed call is not proof of no charge, so verify through usage history rather than trusting the error message
- Delegating discovery to a routing layer makes it easy to stop asking where a number came from, yet the provider behind a capability still governs its accuracy
- The company is very young and publishes neither an address nor an identifiable owner, which is a real counterparty risk for a production dependency
Setup & Integrations
Technical difficulty
Low to moderate, depending on the path. Trying the tool needs nothing: the Playground runs a task from the browser. Connecting an agent needs one install command and an API key. Hosted MCP is the lightest route, with no local process, no Node.js and no package to install, just a URL and a Bearer header. The CLI installs in one line and logs in interactively, while the local MCP server and the OpenClaw plugin require Node.js. Copy-ready prompts let an agent configure itself, and qveris doctor diagnoses key, endpoint and connectivity.
Deployment
Integrations
Supported languages
Behind QVeris
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Is discovery free?
How much does a single call cost?
Is this a subscription?
What do I get for free?
How do I connect my agent?
How do I check whether a failed call was charged?
Where does my query content go?
Is there a minimum age?
Is there a mobile app?
Should you pick QVeris?
QVeris is infrastructure rather than another agent: it exists so that an agent can reach many providers without its builder negotiating, keying and metering each one. On that narrow promise it is well executed. The three-verb protocol is clean, cost is visible before a call rather than after it, the audit surface tells you what you were actually charged, and the client tooling is open source and documented for machines as much as for people. A permanent free tier and USD 19 for 10,000 credits make the cost of finding out very low.
The reservations are about the company, not the product. No postal address appears anywhere on the site, the GDPR is never named, there is no DPA, no subprocessor list and no security certification, and no beneficial owner can be identified. The domain was registered in August 2025 and first archived in January 2026, so this is a young operation, and the reliability figures it advertises cannot be checked against anything public. Query content is forwarded to third-party providers the site does not name.
That combination suits experimentation, prototyping and individual developers well, and sits awkwardly with regulated procurement. If you are building an agent that needs live market data, document parsing or compliance lookups, QVeris is worth an afternoon and a few free credits. If you are putting it under a production workload that carries customer data, get the hosting jurisdiction, the subprocessor list and the corporate identity confirmed in writing first.
- Choosing a selection results in a full page refresh.
- Opens in a new window.