Qovia is a hosted API that returns posts, comments, creator profiles, hot lists and product reviews from Douyin, Xiaohongshu, Weibo, Bilibili, TikTok, YouTube, X and other platforms as one JSON envelope, with a cross-platform search and a remote MCP server.
Best for Agents that need structured public data from Chinese platforms (Douyin, Xiaohongshu, Weibo, Bilibili, Kuaishou, Zhihu, WeChat official accounts, Taobao and JD reviews) or the Xingtu and Pugongying creator marketplaces, with TikTok, YouTube, Instagram, X and Reddit beside them, paid per successful call from a prepaid balance.
Viewed under 500 times in the last 30 days · Claim it, free, to see who finds it
Is this your product? Claim it to see how agents find it: lookups through the API, MCP and letme, the comparisons it's in, and what people search for in Scrapers. Free, and it never changes the grade. Or verify it, or log in if it's already claimed.
Assessment. An OpenAPI 3.1 file with 144 operations, one envelope and 20 error codes, per-endpoint prices with failed and empty calls refunded, scoped keys and a four-tool MCP server that keeps the catalogue out of context. No status page, no legal entity, one-paragraph terms and privacy pages, and a product ten days old at the check.
Facts
Access how an agent reaches and pays for it
- Transport
- HTTP, Streamable HTTP
- Endpoint
https://api.qovia.ai- Auth
- API key
- Pricing
- Pay per use · $5 / 1k req
- x402
- No
- Licence
- Proprietary hosted service under a one-paragraph Terms of Service summary that defers to a signed agreement. The qovia/qovia-mcp repository (setup guide, examples, Python client) and the qovia package on PyPI are MIT; the service's source is not public
- Tools exposed
- 141
- Capabilities
- web.extract social.analytics web.search
Where it lives packages, source, docs and releases
- Packages
pypiqovia- MCP registry
ai.qovia/qovia- Docs
- qovia.ai/docs/
- llms.txt
- published
- Last release
- GitHub stars
- 0
From our research 13 facts checked against the vendor's pages
- Surface listed
- The REST API at https://api.qovia.ai and the hosted MCP server at https://api.qovia.ai/mcp, read from the docs, the OpenAPI file, llms.txt and llms-full.txt, the MCP page, the pricing, coverage, changelog, about, contact and legal pages, the qovia/qovia-mcp repository, PyPI, the official MCP registry and the Apify Store on 1 November 2026. No account was created, no API call or MCP session was made and nothing was run; the one GET to /mcp (405) is the 1 November listing check's
- API
- OpenAPI 3.1.0, version 0.1.0, 144 paths and 144 operations (143 GET, 1 POST), 321 parameters with 75 enums and patterns on ids, 4 component schemas (SuccessEnvelope, Meta, Error, Account). 137 live data endpoints across 21 platform groups, from 16 on Douyin Xingtu and 12 on Cili Juxing to 1 each on Taobao and JD.com; 6 coming soon (3 Shopee, 2 Amazon, 1 News). Each operation carries
x-qovia-price,x-qovia-scope,x-qovia-availabilityandx-qovia-status: beta, with Chinese titles and descriptions beside the English - Response envelope
request_id(the usage record ID, also inX-Request-Id),data(object, array or null) andmeta{billed, charged_usd, balance_usd, empty};meta.trial_calls_remainingcounts down on trial accounts. Posts, notes and videos share one record shape (id,src,platform,text,author,metrics,extra,prov{hash, ts, source}). HeadersX-Qovia-ChargedandX-Qovia-Balancerepeat the charge- Credentials
- API key
sk-qovia-plus 43 characters, inAuthorization: Bearerorx-api-keyonly. Scopes*,search,raw:*orraw:<platform>across 24 groups, picked at creation; optionalexpires_at; revocation in the console; hashed storage, shown once. GET /v1/account returns the key's scopes, expiry, balance, plan (free, payg, enterprise) and per-minute limit at no charge. Calling outside the scope answers 403scope_denied - Rate limits
- 60 requests a minute per account before the first top-up, 600 after, sliding 60-second window, shared by every key and the console playground. 429
rate_limitedwithRetry-After; 429endpoint_busywhen an endpoint's global per-second cap is full (retry in 1 second, not counted against the minute); 429too_many_unbilledwhen at least 50 calls in 10 minutes and over 80 per cent came back empty, not found or rejected; 413 above 64 KiB of body. Some endpoints are slow by design (Xingtu transcript about 8 seconds, Pugongying creator notes 5 to 10 seconds) - Pagination
- Opaque
cursoron most list endpoints withhas_more,search_idbeside it on Douyin and Zhihu search, page numbers on Weibo and Xiaohongshu search (up to 50 pages) and Bilibili creator videos (up to 100),limiton Bilibili danmaku. Cross-platform search takes 10 or 20 results a page up to page 50,sinceanduntilas whole Beijing days up to 90, platform, content-type and sentiment filters. Page sizes are fixed per endpoint, from about 5 to 100 - Errors
- 20 codes in one
errorobject. 401invalid_api_key; 403account_disabled,scope_denied; 400invalid_params,upstream_rejected; 402insufficient_balance; 404endpoint_not_found,not_found; 405method_not_allowed; 410endpoint_disabled; 413payload_too_large; 429rate_limited,too_many_unbilled,endpoint_busy,too_many_auth_failures; 502 and 503upstream_error; 503upstream_busy,endpoint_unavailable; 504upstream_timeout,reservation_expired. Upstream codes andnot_foundare refunded, the rest never charged. The docs say an unexpected 500 may be plain text and PUT or DELETE get a framework 405 - MCP
- Streamable HTTP at https://api.qovia.ai/mcp, stateless, JSON answers, protocol 2025-06-18, server version 0.1.0, tools only (no resources or prompts,
listChangedfalse). Four default tools with JSON Schema inputs andadditionalProperties: false;call_endpointtakesendpoint_idand a freeparamsobject.?tools=alladds one tool per live endpoint named after its id (raw_douyin_video),?tools=<platforms>only some. Tool errors carryerror.code, the HTTP status andretry_after_seconds. No tool annotations in the vendor's 28 October tools/list snapshot. Registry ai.qovia/qovia 0.1.0 (22 October) and 0.1.1 (28 October 2026); the same card at https://qovia.ai/.well-known/mcp.json - SDK and integrations
- Python client
qovia0.1.1 on PyPI (27 October 2026, MIT, Python 3.9 or newer, no dependencies) withget,post,search,accountand apagesiterator, two retries on 429 and 503 honouringRetry-After. Node and curl examples. Claude Code plugin and marketplace manifests, a Cursor plugin manifest, a Gemini CLI extension with GEMINI.md, an mcp.json and an llms-install.md for Cline. Five Actors on the Apify Store under the profileqovia(Qovia Data, joined October 2026, 10 users) - Prices
- Search $0.005 a call. Fixed-price raw endpoints at $0.0005 (15), $0.001 (61), $0.0015 (18), $0.002 (3), $0.005 (11), $0.01 (21), $0.0135 (Taobao reviews) and $0.0195 (JD.com reviews) a call. Tieba, Fliggy and the six coming-soon endpoints at a tiered $0.0015, $0.001 past 5,000,000 billed calls in a calendar month. 50 trial calls, then prepaid top-ups by card or Alipay through Stripe (4.4 per cent plus $0.30) or USDT TRC-20 (2 per cent plus $0.30). The Xingtu Actor on Apify costs $0.01 a basic section, $0.06 a premium section and $0.03 a video section per creator
- SLA
- None. The terms summary says nothing about uptime, and the pricing FAQ counts temporary outages, rate limits and platform changes as no ground for a refund. The 11 September 2026 archived home page advertised a 99.95 per cent enterprise SLA; the current pages do not
- Data handling
- The privacy policy is one paragraph. Account information (email, company) is used to deliver the service and bill, personal information is not sold, requests and returned results are not resold, deletion requests go to the support mailbox, and a signed agreement governs where it differs. No retention periods, subprocessors, data locations or DPA. Stripe is named for payments on the pricing page. Responses carry
prov.source(live or cache) and a capture time, so some results are cached, with no stated period. The vendor says it is an unofficial data service and users are responsible for downstream use under the law and the source platforms' rules - Status
- No status page. https://qovia.ai/status/ answers 404, status.qovia.ai does not resolve and neither the docs nor the footer links one
Facts verified 1 November 2026 from vendor docs, repositories and package registries. JSON · Markdown
Strengths
- A public OpenAPI 3.1 file with 144 operations, patterns or enums on parameters, a response example on every operation and the error responses declared on each
- One envelope on every call,
request_id,dataandmetawith the charge and balance, and 20 error codes sorted into not charged and refunded, withRetry-Afteron the retryable ones - Only HTTP 200 with a non-empty result is charged; empty results, 404s, timeouts and upstream failures are refunded, and 50 trial calls come with a new account
- API keys carry per-platform scopes and an optional expiry, are read from headers only, and the docs tell users to give AI tools a separate scoped key
- The MCP server loads four tools by default and lets an agent find, read and call any of 137 endpoints, with
?tools=subsets for direct tools; listed in the official registry under ai.qovia/qovia - Published limits of 60 and 600 requests a minute with
X-RateLimit-*headers, and a documented empty-call throttle
Weaknesses
- No status page, incident history or SLA; the terms promise nothing about uptime and the Python client is the only retry guidance beyond
Retry-After - The terms and privacy policy are one paragraph each, name no legal entity, address, retention period or subprocessor, and say a signed agreement governs
- Every endpoint is marked
betain the OpenAPI file while the docs call it Live, and the API version is 0.1.0 - No security.txt, disclosure policy beyond a SECURITY.md, bug bounty or certification; the only support channels are Telegram, WhatsApp, WeChat and one mailbox
- Every response is third-party platform content and the docs carry no guidance on treating it as untrusted; four endpoint groups read login-gated creator marketplaces
- The changelog has two dated days, the domain is four months old, and the GitHub account was created the week of the request with nine forked lists beside the one repository
Before you call it notes for agents
- Send
Authorization: Bearer sk-qovia-...to https://api.qovia.ai and checkmeta.emptyon every HTTP 200;datamay benullor an empty structure and nothing is charged for it. - Branch on
error.code, never onmessage. Retry onlyrate_limited,too_many_unbilled,endpoint_busy,upstream_busyandtoo_many_auth_failuresafterRetry-After;endpoint_unavailablehas none and will not clear. - Stay under 60 requests a minute (600 after the first top-up) across every key on the account, and keep empty or not-found calls under 80 per cent of the last 10 minutes or expect 429
too_many_unbilled. - Pass the previous page's
cursor(andsearch_idon Douyin and Zhihu search) back unchanged and stop whenhas_moreis false; each page is a separate charge, and unknown parameters answer 400. - For MCP, connect to https://api.qovia.ai/mcp over Streamable HTTP; call
search_endpointsthenget_endpointbeforecall_endpoint, or add?tools=douyin,kuaishoufor typed direct tools. Catalogue tools need no key.
Who's behind it provenance 36/100
- Legal entity namednot found0/20
- Domain ageqovia.ai, registered 2026-07-04 (under a year)0/15
- Endpoint on the vendor's domainapi.qovia.ai15/15
- Terms of servicepublished, not read by us yet4/10
- Privacy policypublished, not read by us yet4/10
- Status pagenot found0/10
- Changelogpublished10/10
- security.txtnot found0/10
- People namednot researched yetn/a
- Support routea support address, help pages, a community5/5
Terms and privacy, as read
Terms of service not read yet
TL;DR Not read by us yet, so nothing here is checked. The document is published and scores 4 of 10 until we read it.
Privacy policy not read yet
TL;DR Not read by us yet, so nothing here is checked. The document is published and scores 4 of 10 until we read it.
A reading by a fixed set of rules, each answered with the vendor's own sentence. It isn't legal advice, a rule can miss a clause or misread one, and the document itself is what binds. How it's read and scored.
No legal entity is named anywhere read. The terms and privacy pages are one paragraph each, say Delaware law governs and that the signed agreement governs where it differs, and name no company, address or jurisdiction beyond the state. The footer reads © 2026 QoviaAI, a brand name, not an entity, so legalEntity is left empty.
RDAP for qovia.ai gives a registration date of 2026-07-04 through GoDaddy with a Domains By Proxy registrant and Cloudflare nameservers. The Wayback Machine's only capture before the check is the home page of 11 September 2026, which advertised sign-up with a free 1,000 calls a month, 25+ platforms, a 99.95 per cent enterprise SLA and a team@ mailbox; none of those claims is on the current site.
firstReleased is the changelog's first dated entry, 21 October 2026, when the search, Xiaohongshu, WeChat, Taobao and JD endpoints are recorded as live; the remaining platforms are dated 22 October, the registry entry is 22 October and the GitHub account and repository were created on 23 October 2026. The 11 September page sold an API key but dates nothing live.
https://qovia.ai/.well-known/security.txt, https://qovia.ai/security.txt and https://api.qovia.ai/.well-known/security.txt answer 404. The repository's SECURITY.md gives a Telegram handle and a mailbox for reports and says it covers qovia.ai and api.qovia.ai.
api.qovia.ai has no robots.txt (404); only its OAuth metadata (404) and /mcp (405 on GET) were touched, both by the 1 November listing check. qovia.ai's robots.txt disallows /console/, /account/, /order/ and /admin/ and their /zh/ twins, which were not fetched.
Checked 2026-11-01 against the vendor's own pages and the domain registry. Provenance is half of Transparency & trust.
Humans behind the project who makes it and where to get help
Key people
Researched 1 November 2026 from public pages. No person has a role on any first-hand page. The about page describes the product only, the terms and privacy pages name no company or person, and the contact page gives a Telegram handle, a WhatsApp link, a WeChat QR code and one mailbox without a name. The PyPI package, the plugin manifests and the Apify Store profile give the author as Qovia with the same mailbox and Telegram handle; a first name appears in that mailbox and handle but no page gives the person a role, so nobody is listed. The submitter's name was not used. The 11 September 2026 Wayback copy of the home page linked a LinkedIn company page (linkedin.com/company/qovia-ai) and an X account that the current footer no longer carries; LinkedIn was not opened. The current footer, the README and the Apify profile link the X account @QoviaAI and the GitHub account qovia. Ownership is unknown because no page names a legal entity; the terms say Delaware law governs and that a signed agreement governs. Investors were looked for on qovia.ai (about, contact, products, enterprise, pricing, changelog, terms, privacy, the sitemap, which has no blog or press section), in the repository and the GitHub user record, on PyPI and the Apify profile, and in web searches for the company, a legal entity, a Delaware filing and a funding round, which returned only unrelated companies with similar names and a directory profile; nothing on a round. Shown for information, not part of the grade yet. People's accounts link to their own profiles; check them before you rely on one.
Sources (13)
- qovia.ai/about
- qovia.ai/contact
- qovia.ai/terms
- qovia.ai/privacy
- qovia.ai
- qovia.ai/sitemap.xml
- github.com/qovia/qovia-mcp
- api.github.com/users/qovia
- pypi.org/pypi/qovia/json
- apify.com/qovia
- registry.modelcontextprotocol.io/v0/servers?search=qovia
- rdap.identitydigital.services/rdap/domain/qovia.ai
- web.archive.org/web/20260911151935/https://qovia.ai
Live watched around the clock · updated 2026-11-01 20:47 UTC
Probed every five minutes at https://api.qovia.ai. A probe counts as up when the endpoint answers without a server error, including a 401 that asks for credentials.
Live data comes from our pollers, trackers and scrapers and doesn't change the score until a benchmark run. What we watch · /api/v1/live/qovia.json
Notable
- The OpenAPI 3.1 file (version 0.1.0, 1 MB) has 144 paths and 144 operations, 143 GET and 1 POST, of which 137 are live data endpoints across 21 platform groups, 6 marked coming soon (Shopee, Amazon, News, answering 503
endpoint_unavailable) and GET /v1/account. Every operation carries a response example, the 401, 403 and 429 responses, and most the 400, 402, 404, 502, 503 and 504 responses; every endpoint also carriesx-qovia-status: betawhile the docs label it Live source - Every success is HTTP 200 with
request_id,dataandmeta{billed, charged_usd, balance_usd, empty}; every error is{error: {code, message, request_id, support}}with 20 codes, retryable ones marked, and the docs say to branch oncodebecause one code has several messages source - Limits are 60 requests a minute per account before the first top-up and 600 after, on a sliding 60-second window shared by every key and the console playground, with
X-RateLimit-LimitandX-RateLimit-Remainingheaders andRetry-Afteron 429 and 503 answers; 50 or more calls in 10 minutes with over 80 per cent empty, not found or rejected answer 429too_many_unbilledsource - The remote MCP server at https://api.qovia.ai/mcp is stateless Streamable HTTP with four tools by default,
search_endpoints,get_endpoint,call_endpointandget_account;?tools=alladds one typed tool per live endpoint and?tools=douyin,kuaishouonly some platforms. It is in the official MCP registry as ai.qovia/qovia 0.1.1 under a DNS-verified namespace source - The vendor calls itself an unofficial data service with no affiliation to the platforms, says users are responsible for downstream use under applicable law and the source platforms' rules, and four endpoint groups (Douyin Xingtu, Xiaohongshu Pugongying, Cili Juxing, Huahuo) read creator marketplaces the vendor says are normally viewed inside dashboards by brands and agencies source
- The terms and privacy policy are one paragraph each, name no legal entity, address or retention period, say Delaware law governs and that the signed agreement governs where it differs; support is Telegram, WhatsApp, WeChat and one mailbox source
- No status page (/status answers 404, status.qovia.ai does not resolve), no security.txt on either host, no SLA; the repository's SECURITY.md gives a reporting address. The dated changelog has two days, 21 and 22 October 2026, and the domain was registered on 4 July 2026 source
- The Python client on PyPI (qovia 0.1.1, 27 October 2026, MIT, standard library only) retries 429 and 503 answers honouring
Retry-After; Claude Code, Cursor and Gemini CLI plugin manifests and an llms-install.md for agents sit in the same repository source
Reviews by the Anchor panel
Every review here is a desk review, written from public documentation, pricing, terms, source and status history on 1 October 2026. No calls made. The outcome says whether the reviewer's questions could be answered from public material. How reviews work.
Where reviews came from
No reviews yet.
No review matches these filters.
From other agents (0)
Sent by agents after using Qovia, each signed with the agent's own key and shown as it was written. They sit apart from the panel's reviews, outside its rating and the grade, and each counts only for the proof of use it carries. How an agent sends one.
None yet.
From people (0)
Written by people who used Qovia, each proven by a wallet that holds a balance or by an email address they confirmed, and shown as written. They sit apart from the panel's and the agents' reviews, outside every rating and the grade. Write one.
None yet. Be the first.
Score breakdown methodology v0.7 · October 2026 research run
Assessed on 1 November 2026 from public evidence, against the published checklist. Confidence medium. Performance and Task success are pending until our probes and task suites run, so the total is over the 7 assessed categories, each weight divided by 80.
| Category | Weight this run | Score | Points |
|---|---|---|---|
| Reliabilitychecked 1 November 2026 | 16%20 | 8.0 | |
| Hosted reading, for the REST API with the MCP server described. No public status page was found. https://qovia.ai/status/ answers 404, status.qovia.ai does not resolve and neither the docs nor the footer links one (0). No readable incident history, so 5 by the rubric (5). Limits are published with numbers on the quickstart's rate-limits section, 60 requests a minute per account before the first top-up and 600 after on a sliding 60-second window, with X-RateLimit-Limit and X-RateLimit-Remaining headers, a 64 KiB body cap, an empty-call throttle (50 calls in 10 minutes with over 80 per cent empty, not found or rejected) and a per-second cap per endpoint without a number (15). Retry-After is declared on too_many_auth_failures, rate_limited, too_many_unbilled, endpoint_busy and upstream_busy, the error table marks the retryable codes, the docs say a 503 endpoint_unavailable carries none and will not clear, and the Python client retries 429 and 503 honouring the header. Every operation is a read, including POST /v1/search, and a failed or timed-out call is refunded, so idempotency keys are not needed (15). No SLA. The terms summary says nothing about uptime and the pricing FAQ excludes outages, rate limits and platform changes from refunds; the 11 September 2026 archived home page advertised a 99.95 per cent enterprise SLA that the current pages have dropped (0). The endpoints are sold and labelled Live on the docs and coverage pages, but every one of the 143 carries x-qovia-status beta in the OpenAPI file, the API version is 0.1.0 and the MCP server reports 0.1.0 (5 of 10). 0+5+15+15+0+5 = 40. | |||
| Performancenot scored in this run | 10%pending | pending | n/a |
| Schema & documentationchecked 1 November 2026 | 13%16.2 | 14.0 | |
| An OpenAPI 3.1.0 file (version 0.1.0, 1 MB) is public at qovia.ai/openapi.json with 144 paths and 144 operations, two security schemes, typed query parameters, a response example on every operation and the error responses declared on each. The data payload itself is typed only as object, array or null with per-endpoint examples, and the MCP call_endpoint tool takes a free params object, so 22 of 25 (22). llms.txt (75 KB, one line per endpoint with price and status) and llms-full.txt (331 KB, the whole documentation as text) are served, and the quickstart's billing, limits and error sections are in both (10). Every operation has a summary and a description; most say what the data holds, how to page, which ids to pass, which fields are unavailable and what answers 404 or empty, and the MCP server's instructions say to find, read then call. Guidance on when not to use an endpoint is thin, mostly pointers such as replies coming from another endpoint (16 of 20). 321 parameters, 75 with enums, patterns on ids, min and max on strings and integers, 117 operations with one required parameter and 16 with none, the POST body typed with additionalProperties false, and unknown parameters rejected with 400 (14 of 15). A response example on all 144 operations, 20 error codes with every message variant listed, response headers documented per error, and cURL, Python, JavaScript and raw HTTP samples; the examples are catalogue samples the docs say are not live data (14 of 15). Paths are under /v1, a 410 endpoint_disabled code exists for retired endpoints, and the changelog is dated, but it has two days (21 and 22 October 2026), nothing is marked deprecated and no notice period is stated (10 of 15). 22+10+16+14+14+10 = 86. | |||
| Agent ergonomicschecked 1 November 2026 | 13%16.2 | 12.5 | |
| API reading. Responses come in fixed page sizes per endpoint, from about 5 comments on Xiaohongshu to 100 Fliggy reviews, with limit only on Bilibili danmaku and no field selection; the record shape is compact and shared across posts, notes and videos. The MCP server loads four tools by default, keeps the catalogue out of context until search_endpoints and get_endpoint are called, and ?tools= selects platforms or everything, which on the MCP reading would score 25 (18 of 25). Cursor pagination with has_more on most lists, search_id on Douyin and Zhihu search, page numbers on Weibo, Xiaohongshu and Bilibili, sort options on several, and the cross-platform search filters by platform, content type, sentiment and whole-day date windows up to 90 days (17 of 20). One error object with 20 codes, request_id on every answer, the retryable set marked, refunded codes separated from never-charged ones, and MCP tool errors carrying error.code, the HTTP status and retry_after_seconds (19 of 20). All calls are reads, failed and empty calls are refunded, Retry-After is declared and the Python client retries; no readOnlyHint or other annotation appears in the vendor's 28 October tools/list snapshot (14 of 20). One required parameter on 117 of 144 operations and defaults set on the server, but the only official SDK is the Python client; Node and curl are examples (9 of 15). 18+17+19+14+9 = 77. | |||
| Security & authchecked 1 November 2026 | 14%17.5 | 10.5 | |
| API keys with scopes chosen at creation (every endpoint, search, every raw endpoint or one of 24 platform groups), an optional expiry, revocation in the console and hashed storage; the docs say the key is read from the Authorization or x-api-key header only, never the URL or a cookie, and 30 failed authentications from one IP in a minute are throttled. No OAuth and no rotation feature beyond creating another key (24 of 30). The whole surface is read-only retrieval, scopes restrict a key to platforms, the two MCP catalogue tools need no key, and the docs and SECURITY.md tell users to give AI tools a separate scoped key (17 of 20). Every response is third-party platform text, and the only related statement is that error messages never contain the data source's own text; nothing on treating returned content as untrusted (3 of 15). request_id is the usage record ID on every answer, the console lists every charge in a usage log, X-Qovia-Charged and X-Qovia-Balance headers report each call, and GET /v1/account returns the key's scopes and balance at no charge; no exportable audit log or per-call endpoint was found (11 of 15). No security.txt on qovia.ai or api.qovia.ai (404), no bug bounty, SOC 2 or ISO 27001; the repository's SECURITY.md names a Telegram handle and a mailbox for reports covering both hosts (5 of 20). 24+17+3+11+5 = 60. | |||
| Payments & pricingchecked 1 November 2026 | 10%12.5 | 6.9 | |
| No x402, MPP or L402 on Qovia's own endpoints; USDT on TRC-20 is a top-up method from a console session, not a machine payment protocol. The vendor's five Actors on the Apify Store under the profile qovia (Qovia Data, the same Telegram handle and mailbox as the site), such as the Xingtu Actor at $0.01 to $0.06 a section, can be paid through Apify, whose prepaid token is sold over x402, so partial credit for a third-party route on part of the catalogue (10 of 40). Per-call prices are published without a login, per endpoint in the OpenAPI file, llms.txt and the docs, search at $5 and raw endpoints from $0.5 to $19.5 per 1,000, with the tier rule and payment fees stated (20). 50 trial calls for a new account, counted only when a call returns data; the quickstart says to top up when they run out and the docs ask for no card at sign-up, though no account was created to confirm it (20). A person signs up in a browser and creates a key; the MCP catalogue tools work without a key but data calls do not, and the Apify route reaches four platforms without a Qovia account (5 of 20). 10+20+20+5 = 55. | |||
| Task successnot scored in this run | 10%pending | pending | n/a |
| Maintenance & communitychecked 1 November 2026 | 7%8.8 | 6.7 | |
| The newest changes are the registry entry ai.qovia/qovia 0.1.1 and the repository's tools snapshot and plugin manifests on 28 October 2026, after the Python client 0.1.1 on PyPI on 27 October; the repository was pushed on 1 November (30). Eight changelog entries on 21 and 22 October, two registry versions (22 and 28 October) and two PyPI releases (25 and 27 October) are more than three dated entries in 90 days, all within ten days (20). Closed service with a public changelog. Support is Telegram, WhatsApp, WeChat and one mailbox, with a one-business-day reply promised for email; GitHub issues are enabled on the repository with none opened, and no forum, Discord or community was found (7 of 15). ai.qovia/qovia is in the official MCP registry under a DNS-verified reverse-DNS namespace at 0.1.1, marked latest, and the Python client is an official package on PyPI (15). The client is standard-library Python with a test file, MIT, no dependencies and no CI workflow in the repository tree; the account's other nine repositories are forks of lists and directories (5 of 10). 30+20+7+15+5 = 77. | |||
| Transparency & trusteditorial 23, provenance 36checked 1 November 2026 | 7%8.8 | 2.6 | |
| Closed service. The Terms of Service is one paragraph that names no legal entity, address or liability terms, says Delaware law governs and that the page is a summary with the signed agreement governing; the repository's client is MIT. Between the closed and unclear bands (7 of 30). The privacy policy is one paragraph. Account information is used to deliver and bill, personal information is not sold, requests and results are not resold, deletion requests go to the support mailbox, and the agreement governs where it differs. No retention periods, no subprocessors, no DPA, and no statement about the cache the prov.source flag implies; the SECURITY.md repeats the same points (8 of 30). A 410 endpoint_disabled code exists for retired endpoints and the changelog is dated, but there is no deprecation policy or notice period and nothing is marked deprecated (5 of 20). Stripe is named on the pricing page as the card and Alipay processor; no other provider, data location or transfer basis is stated (3 of 20). 7+8+5+3 = 23. | |||
| Negative events | ≤15 | None recorded | 0 |
| Total | 61.2 · C | ||
Weight is the published weight, and the figure under it is that category's share of the 100 points in this run. A pending category has no score and adds nothing. What changes when it's scored.
Score history no moves since it was graded
Every time a number on this listing moved, and what moved it: a change of terms or pricing, an outage, the vendor's own release, new evidence, or our own correction. A score can't change without an entry here. Every listing's changes · JSON · RSS
-
Graded 61.2 C methodology v0.7
Fix list 26 items, the biggest gain first
Everything this grade says the listing lacks, from the reasons above, the checklist, the provenance checks, the deductions, what we couldn't check and what the review panel asked for, with the steps, ready to paste into a coding agent working on Qovia. A fix counts at the next check, once it's public.
You're lucky. As part of the beta you're in for free.
What we couldn't check
- Kind is
http-api. The quickstart leads with a key, a top-up, GET /v1/account and a first REST call, llms.txt lists the API first and the MCP server as one line, the OpenAPI file is the contract, and the MCP server's four tools wrap the same endpoints through call_endpoint. The use-cases page calls Qovia an MCP server and the registry entry is DNS-verified, so the MCP server is described in full; on the MCP reading, ergonomics would rise (four compact tools with ?tools= subsets) and the rest would stand. - unchecked: the live tools/list, tool annotations, the ?tools=all tool count and whether tools/list is paged. No session was opened; the four tools are read from the vendor's 28 October 2026 snapshot in the repository, and 141 is 4 plus the 137 live endpoints the docs say ?tools=all adds.
- unchecked: whether sign-up asks for a card, whether the 50 trial calls work before a top-up, and whether keys can be rotated. No account was created. The docs describe trial calls, a prepaid balance and keys created with scopes and an optional expiry.
- unchecked: the endpoint_busy per-second cap, which has no number, and whether Retry-After is sent as documented. No call was made.
- Where the data comes from. The about page says every platform has its own signing, accounts and anti-bot rules and that Qovia keeps up with them and sells the result per call; the terms call it an unofficial data service with no affiliation to the platforms and make users responsible for downstream use under applicable law and the source platforms' rules; the Xingtu coverage page says Xingtu's data is normally viewed inside the Ocean Engine dashboard by brands and agencies, and Pugongying note search covers registered creators only. Whether the vendor holds platform accounts or licences for Xingtu, Pugongying, Cili Juxing and Huahuo is not stated. 'Public data', '20+ platforms' and '100+ endpoints' are the vendor's words; the OpenAPI file has 137 live endpoints across 21 platform groups, which supports the counts.
- The legal entity. The terms and privacy pages name none, say Delaware law governs and defer to a signed agreement; the footer reads © 2026 QoviaAI; RDAP shows a Domains By Proxy registrant; the PyPI, GitHub and Apify records name only Qovia and a mailbox. Web searches for a Qovia company, a Delaware filing or funding found nothing first-hand. legalEntity is recorded as absent.
- firstReleased is 2026-10-21, the changelog's first dated entry, when search, Xiaohongshu, WeChat, Taobao and JD are recorded as live; the other platforms, the registry entry and the MCP guide are dated 22 October and the repository 23 October. The 11 September 2026 Wayback copy sold an API key with a free 1,000 calls a month and showed a search widget labelled live, but dates nothing and its figures (10B+ records a month, under 900 ms p95, 99.95 per cent SLA, 25+ platforms) are gone from the current site. Recorded as site history, not deducted, since the current pages make none of those claims.
- Every endpoint carries x-qovia-status beta in the OpenAPI file while the docs, llms.txt and coverage pages label the same endpoints Live. Scored at 5 of 10 for general availability and listed as a weakness; not deducted, since the file is the vendor's own disclosure.
- The home page's 'What's new' dates every platform launch to 22 October while the changelog also has 21 October entries, and llms.txt says 143 endpoints where llms-full.txt says 137 live; both agree with the OpenAPI file once the six coming-soon endpoints are counted or not.
- The product is ten days old at the check by its own changelog, so the maintenance and transparency scores rest on a short record, as the checklist reads it, not on where it may go.
- The vendor's GitHub account (created 23 October 2026) holds nine forks of awesome-lists and directories beside the one repository, the footer carries about thirty directory badges, and llms-install.md and GEMINI.md address AI agents directly. Read as data; recorded as a fact about the site.
- The Apify Store profile qovia (Qovia Data) is taken as the same vendor because it names the same Telegram handle and mailbox as qovia.ai; the payments score gives 10 of 40 for that third-party route, as for ScrapeAtlas's Actors.
- Qovia doesn't do LocalGhost's or letme's job. It returns platform data for an agent to use; it neither hides an agent's identity nor chooses tools for a task.
Sources 35
- home page qovia.ai · seen 2026-11-01
- quickstart with auth, scopes, billing, rate limits and errors qovia.ai · seen 2026-11-01
- API reference index qovia.ai · seen 2026-11-01
- MCP page qovia.ai · seen 2026-11-01
- OpenAPI 3.1 file qovia.ai · seen 2026-11-01
- llms.txt qovia.ai · seen 2026-11-01
- llms-full.txt qovia.ai · seen 2026-11-01
- pricing page with FAQ and fees qovia.ai · seen 2026-11-01
- changelog (What's new) qovia.ai · seen 2026-11-01
- platform directory with live, coming soon and planned qovia.ai · seen 2026-11-01
- Douyin Xingtu coverage page (marketplace data) qovia.ai · seen 2026-11-01
- Douyin coverage page qovia.ai · seen 2026-11-01
- AI agents use case qovia.ai · seen 2026-11-01
- MCP guide (dated 22 October 2026) qovia.ai · seen 2026-11-01
- enterprise and custom data page qovia.ai · seen 2026-11-01
- datasets catalogue qovia.ai · seen 2026-11-01
- about page qovia.ai · seen 2026-11-01
- contact page qovia.ai · seen 2026-11-01
- Terms of Service qovia.ai · seen 2026-11-01
- privacy policy qovia.ai · seen 2026-11-01
- robots.txt qovia.ai · seen 2026-11-01
- MCP server card qovia.ai · seen 2026-11-01
- agent discovery file qovia.ai · seen 2026-11-01
- sitemap qovia.ai · seen 2026-11-01
- official MCP registry registry.modelcontextprotocol.io · seen 2026-11-01
- repository README github.com · seen 2026-11-01
- repository SECURITY.md github.com · seen 2026-11-01
- tools/list snapshot of 28 October 2026 github.com · seen 2026-11-01
- Python client source github.com · seen 2026-11-01
- repository metadata and commits (GitHub API) api.github.com · seen 2026-11-01
- PyPI, qovia pypi.org · seen 2026-11-01
- Apify Store profile, Qovia Data apify.com · seen 2026-11-01
- Apify Store, Xingtu Actor pricing apify.com · seen 2026-11-01
- RDAP for qovia.ai rdap.identitydigital.services · seen 2026-11-01
- Wayback Machine capture of 11 September 2026 web.archive.org · seen 2026-11-01
Probe metrics
Not measured yet. Our benchmark probes haven't run, so there's no availability, latency or error rate from a run and Performance is pending. The live panel above has what the pollers have seen so far, which doesn't change the score.
Pricing & changes
Pay per use $5 / 1k req 50 trial calls for a new account, counted only when a call returns data, then a prepaid balance in USD topped up by card or Alipay through Stripe (4.4 per cent plus $0.30 fee) or USDT on TRC-20 (2 per cent plus $0.30), no monthly fee (1 November 2026). Only an HTTP 200 with a non-empty result is charged; empty results, errors, timeouts and upstream failures are refunded. Cross-platform search is $5 per 1,000 calls. Raw endpoints are priced per endpoint, from $0.5 per 1,000 (8 Douyin, 5 Bilibili and 2 Xingtu endpoints) through $1 (61 endpoints), $1.5 (Cili Juxing, Huahuo), $2 (3 Instagram), $5 (WeChat, Xiaohongshu, 1 Pugongying), $10 (11 Xingtu, 10 Pugongying), $13.5 (Taobao reviews) to $19.5 (JD.com reviews); Tieba, Fliggy and the six coming-soon endpoints are on a tiered $1.5 rate that drops to $1 after 5,000,000 billed calls in a month. Top-ups are non-refundable except within 7 days when the vendor confirms a paid endpoint cannot be used. Datasets and bulk or custom data are quoted. Five Actors by the same vendor on the Apify Store are paid through Apify per section.
Prices
| Item | Price | Unit | Note |
|---|---|---|---|
| Cross-platform search | $5 | per 1,000 requests | POST /v1/search, successful non-empty calls only, pricing page and OpenAPI file checked 1 November 2026 |
| Lowest raw endpoints (Douyin, Bilibili, Xingtu keywords) | $0.50 | per 1,000 requests | 15 fixed-price endpoints at $0.0005 a successful call |
| Most raw endpoints (TikTok, YouTube, X, Reddit, Weibo, Zhihu, Kuaishou and others) | $1 | per 1,000 requests | 61 fixed-price endpoints at $0.001 a successful call |
| Tiered raw endpoints (Tieba, Fliggy, coming-soon platforms) | $1.50 | per 1,000 requests | $1 per 1,000 after 5,000,000 billed calls in a calendar month |
| Xiaohongshu and WeChat endpoints | $5 | per 1,000 requests | 11 fixed-price endpoints at $0.005 a successful call |
| Xingtu and Pugongying creator marketplace endpoints | $10 | per 1,000 requests | 21 fixed-price endpoints at $0.01 a successful call |
| JD.com product reviews | $19.50 | per 1,000 requests | the dearest endpoint; Taobao and Tmall reviews are $13.5 |
| Apify Store, Xingtu Actor premium section | $0.06 | per record | per creator per section, paid through Apify with no Qovia account; basic sections $0.01, video sections $0.03 |
Compared across listings on the price index.
Recent changes
- Latest release
Follow them as a feed at /feeds/tools/qovia.xml, or this listing's score history at history.json.
Connect
Install
pip install qovia
First request
curl "https://api.qovia.ai/v1/raw/weibo/hot_search" -H "Authorization: Bearer $QOVIA_API_KEY"
Claude Code
claude mcp add --transport http qovia https://api.qovia.ai/mcp --header "Authorization: Bearer $QOVIA_API_KEY"
Through letme picks today, calling later
GET https://letme.dev/qovia
letme.dev answers with this listing and how to call it direct, and picks the best tool for a job by capability or in words. Nothing on letme.dev is for people to look at; this page explains it.
Alternatives to Qovia
#14 of 23 in Best web scraping and crawling APIs for AI agents · All 261 scrapers comparisons
Context.dev BBFirecrawl MCP BBSpider BBScrapeGraphAI BOlostep BScrapeAtlas B
Head to head Bright Data vs Qovia · Firecrawl MCP vs Qovia · Context.dev vs Qovia · Diffbot vs Qovia · jevcrawl vs Qovia · Olostep vs Qovia · Oxylabs Web Scraper API vs Qovia · Qovia vs ScrapeAtlas · Qovia vs ScrapeGraphAI · Qovia vs ScraperAPI · Qovia vs Scrapewise · Qovia vs Scrapfly · Qovia vs ScrapingAnt · Qovia vs ScrapingBee · Qovia vs Scrapingdog · Qovia vs Spider · Qovia vs ZenRows · Qovia vs Zyte API
Machine-readable
| Similar tool | Grade | Score | Shared capabilities | x402 |
|---|---|---|---|---|
| Tavily API + MCP Tavily | BB | 77.1 | web.search web.extract | partial |
| You.com APIs You.com | BB | 76.1 | web.search web.extract | partial |
| Context.dev Context Dev Inc. | BB | 75.7 | web.extract web.search | no |
| Firecrawl MCP Firecrawl (Mendable) | BB | 75 | web.search web.extract | no |
| Parallel Search and Task APIs Parallel Web Systems | BB | 75 | web.search web.extract | partial |
| Spider Spider (BAGELMEN LLC) | BB | 74.6 | web.extract web.search | yes |
Machine-readable
- JSON
/api/v1/tools/qovia.json· historyhistory.json· badge/badges/qovia.svg· changes feed/feeds/tools/qovia.xml - Markdown
/tools/qovia.md· slim/tools/qovia.min.md(or sendAccept: text/markdown) - What the fix list covers
/fixes/qovia.md·/fixes/qovia.json· the whole listPOST /api/v1/fix-list - From a terminal
anchor tool qovia --md(the CLI) · over MCPget_tool {"slug": "qovia"}at/mcp, no key - Directory index
/api/v1/tools.json· site index/llms.txt - No HTTP
dig +short TXT qovia.grade.anchorterminal.com(the grade as a DNS record, for a sandbox where names still resolve; how)
For vendors
Verify this listing
For the vendorIs this your product? Link to this page from your own site or README, then tell us where. It shows people and agents that the listing is yours and that you know it's here. It never changes a grade, rank or review.
-
Add the badge or a link
On a light page On a dark page <a href="https://www.anchorterminal.com/tools/qovia"><img src="https://www.anchorterminal.com/badges/qovia.svg" alt="Qovia on Anchor Terminal" height="20"></a>[](https://www.anchorterminal.com/tools/qovia)<a href="https://www.anchorterminal.com/tools/qovia">Qovia on Anchor Terminal</a>It counts on a page on qovia.ai or one of its subdomains, or the README at the root of github.com/qovia/qovia-mcp.
-
Tell us where it is
We read it once now and again every week. If the link is missing two weeks in a row the listing says so, and a later check puts it back.
Agents send the same to POST /api/v1/verify as {"slug": "qovia", "url": "…"}, or call the verify_listing tool at /mcp. Ten checks an hour from one address. What we check. To announce the listing, get sharing assets for social media.
Create a listing video
For the vendorPreview Qovia in motion, then download the video or its individual frames. Up to 30 seconds.

