Vectorize
Embedding API built for AI agents. No API keys, no accounts. Send payment, get vectors. x402 micropayments on Solana.
Vectorize Your Data
Upload chunked documents or raw text. Select your tier and target database format. Get production-ready vectors.
Built for AI Agents
No API keys. No accounts. No OAuth. Just pay and use. Programmatic access designed for autonomous systems.
Zero Config
No API keys to manage. No developer accounts. Discover pricing at GET / and start embedding.
Pay Per Request
USDC on Solana. Sub-second finality. Include TX signature in X-PAYMENT header.
Machine Readable
JSON in, JSON out. Consistent error codes. Built for programmatic access by autonomous agents.
Agent Integration Example
# 1. Discover pricing
GET https://api.vectorize.cc/
# 2. Estimate cost
POST https://api.vectorize.cc/estimate
{"chunks": ["text1", "text2"], "tier": "pro"}
# 3. Pay (Solana USDC transfer to payTo address)
# 4. Vectorize with payment proof
POST https://api.vectorize.cc/vectorize/pro
X-PAYMENT: <solana-tx-signature>
{"chunks": ["text1", "text2"], "output_format": "pinecone"}Models By Tier
Each tier uses a specific model optimized for that price point. No configuration needed.
all-MiniLM-L6-v2
Script & Agent TiersMicrosoft's distilled sentence-transformer trained on 1B+ sentence pairs from diverse sources. Runs entirely on-device with zero external API calls—your data never leaves the server. 6-layer architecture delivers 5x faster inference than BERT-base while maintaining 90% of its accuracy. Ideal for high-throughput workloads where speed and privacy matter more than maximum precision.
- Startups — MVP prototyping, validating RAG concepts before scaling
- E-commerce — Product search, recommendation engines, catalog matching
- Internal Tools — Company wikis, Slack/Discord bots, FAQ systems
- Privacy-First — Healthcare prototypes, on-prem deployments, GDPR compliance
- • Semantic similarity search
- • Question-answer matching
- • Duplicate detection
- • Clustering & classification
- • Paraphrase identification
- • Dimensions: 384
- • Max tokens: 512
- • Latency: ~5ms/chunk
- • Provider: Local (sentence-transformers)
- • Languages: English-optimized
text-embedding-3-small
Pro TierOpenAI's production-grade embedding model, released January 2024. Achieves 62.3% on MTEB benchmark while being 5x cheaper than ada-002. Native support for 100+ languages with strong cross-lingual retrieval. Handles documents up to 8,191 tokens—perfect for long-form content without chunking overhead. The sweet spot between cost, speed, and quality for most production workloads.
- SaaS Products — Help centers, documentation search, in-app assistants
- Customer Support — Ticket routing, knowledge base search, agent assist
- Content Platforms — Article recommendations, content discovery, tagging
- Enterprise Search — Confluence, Notion, Google Drive indexing
- • Multilingual semantic search
- • Long document embedding
- • Cross-lingual retrieval
- • Code search & documentation
- • RAG with hybrid search
- • Dimensions: 1,536
- • Max tokens: 8,191
- • Latency: ~50ms/chunk
- • MTEB score: 62.3%
- • Languages: 100+
text-embedding-3-large
Pro+ TierOpenAI's highest-fidelity embedding model with 3,072 dimensions. Achieves 64.6% on MTEB—top-tier performance for nuanced semantic understanding. Excels at distinguishing subtle differences in meaning, technical terminology, and domain-specific jargon. The extra dimensionality captures relationships that smaller models miss, critical for high-stakes retrieval where precision matters.
- Legal — Contract analysis, case law search, due diligence, compliance review
- Medical/Biotech — Clinical trial data, research papers, drug interaction databases
- Financial — SEC filings, earnings calls, risk analysis, regulatory documents
- Research — Academic papers, patents, technical specifications, scientific literature
- • Fine-grained semantic matching
- • Technical jargon understanding
- • Multi-hop reasoning support
- • Disambiguation of similar terms
- • Complex query understanding
- • Dimensions: 3,072
- • Max tokens: 8,191
- • Latency: ~80ms/chunk
- • MTEB score: 64.6%
- • Languages: 100+
Your Database, Your Choice
Export vectors in the exact format your database expects. Zero conversion needed.