Verification API for AI apps and agents
LLMs make things up. We check their work against real sources and return citations with confidence scores. Works with OpenAI, Claude, Gemini, or whatever model you're building on.
Proof for every damn thing
Features
Built to verify, extract, and cite. So your AI apps and agents work with facts, not hallucinations.
Pinpointed source extraction
Pull the exact passage from each source that backs up a claim. Ready for research, due diligence, or agent pipelines.
Reverse citation search
Find out who's citing or supporting your content across the web. Useful for competitive intel and market research.
Citation classification
Sort every citation by source type: journals, news, government records, industry reports. Know where each fact comes from.
Source Integrity
Evidence Strength
96%
Fact Consistency
Domain Credibility
Credibility Scoring
Score how much each source can be trusted before your team or AI agent acts on it.
Bring your own data
Upload your own data. Internal reports, knowledge bases, business docs. It stays encrypted and indexed for your team only.
Security & compliance
Built for regulated industries. Encryption at rest, we don't store your queries, and we're SOC 2 ready.
How it works
How we lock in credibility
Trusted data ingestion
We pull in and keep updating verified public and institutional sources. Academic journals, business publications, government records, industry reports.
Extract and index
Each document gets embedded, indexed, and tagged by domain, publisher, and type. This builds a structured knowledge base your team and agents can actually rely on.
Rank by credibility
Results get scored across relevance, credibility, recency, and domain match. Only the strongest references make it through to your response.
Deliver verifiable proof
You get back structured JSON with citations and reliability scores. Plug it into your LangChain agent, research report, or any workflow that needs facts it can point to.
Pricing
Simple, transparent pricing
Start free, scale as you grow. All plans include full API access and our credit-based billing system.
How credits work
Each claim verification uses credits: 2 credits for citation search, +1 credit for stance analysis, +1 credit for verdict. Full verification = 4 credits. Save credits by skipping stance or verdict when you don't need them.
Free
$0
foreverFor individuals, analysts, and researchers exploring AI-powered fact-checking.
50 credits per month
Full verification (citation + stance + verdict)
Access to web sources
Basic API access
Community support
JSON export
Builder
$20
/monthFor developers building AI apps and agents with citation verification.
500 credits per month
Everything in Free
$0.03/credit for overages
Full API & SDK access
Streaming SSE support
Priority support
Webhook notifications
JSON + CSV export
Enterprise
Custom
For consulting firms, financial institutions, and organizations with high-volume verification and compliance needs.
10,000+ credits per month
Everything in Builder
Bring your own organizational data
Granular file citations
Data integrations (Google Drive, SharePoint, S3)
Custom fields & citation logic
SSO & SAML authentication
Custom SLA & uptime guarantee
Dedicated account manager
FAQ
We're here to help
Webcite is a verification API that fact-checks claims and pulls citations from real sources. Send any claim from your AI agent, whether it runs on OpenAI, Claude, Gemini, Llama, or something you fine-tuned yourself. We search journals, news, and government records to find evidence that supports or contradicts it. Integrate through our REST API or real-time streaming (SSE). Most teams drop it into their LangChain or LlamaIndex pipeline as a verification step.
Each verification includes three components: 1. Citation Extraction - Find and extract relevant sources with credibility scores 2. Stance Analysis - Determine if sources support or contradict the claim 3. Verdict - Get an overall verification result (supported, contradicted, mixed, unverifiable) You can skip stance or verdict to save credits when you only need citation extraction.
Credits are consumed per operation: citation extraction costs 2 credits, stance analysis costs 1 credit, and verdict generation costs 1 credit. A full verification (all three) costs 4 credits total. You can save credits by skipping stance or verdict when you don't need them.
Yes. Webcite fits right into AI agent pipelines and business operations. Add a verification step after your LLM generates output and catch hallucinations before they reach anyone who matters. Works with OpenAI, Claude, Gemini, Llama, and any model that produces text. People use it for research that needs cited sources, financial due diligence, content pipelines that need fact-checked output, compliance workflows, and market research where you can't afford to be wrong.
Generate an API key from your dashboard and make REST calls to our endpoints. We support both synchronous responses and real-time streaming via Server-Sent Events (SSE) for live updates as citations are extracted. Check our API docs for code examples in JavaScript, Python, and cURL. Most developers integrate WebCite as a verification step in their AI agent or research pipeline.
On the Free plan, you'll need to wait until the next month. On the Builder plan ($20/month), you can continue using the API with usage-based billing at $0.03 per additional credit. Enterprise plans include custom allocations for high-volume research and compliance workflows.
Enterprise includes: 10,000+ credits/month, bring your own organizational data, data integrations (Google Drive, SharePoint, S3), custom citation extraction logic, SSO/SAML authentication, dedicated account manager, and custom SLAs. Built for organizations in consulting, finance, healthcare, and legal that run AI agents at scale with compliance and audit requirements.
Yes. All data is encrypted in transit and at rest. We don't store your verification queries beyond processing. Enterprise customers get their own data sources with granular access controls, which matters a lot if you're in finance, healthcare, or legal.
Articles
