One API for embeddings.
OpenAI, Gemini, Voyage & local Llama.
One API call. Any provider. Built-in caching, usage tracking, and token overflow handling — so you can focus on your product, not your embedding pipeline.
Route requests across providers with a single schema. Switch providers or models without touching client code.
Need a model that’s not listed? Request it and we’ll make it available on the Llama provider.
Reduce cost and latency with deterministic cache keys.
When an embedding is already cached, external provider calls are skipped entirely — so you only pay once per unique input.
Add your provider keys once in the dashboard, then use a single vectors.space key in your app — no secrets scattered across services.
Dashboard
Usage, performance, and request insights at a glance.
Request Distribution
API calls by model
Token Usage
Last 12 months
Cache Hit Rate
Last 30 days
Provider Key Vault
Store your keys for OpenAI, Gemini, and Voyage securely. Our built-in Llama provider requires no configuration.
Service Keys
Create and rotate vectors.space API keys for your applications and environments.
Detailed Request Logs
Inspect searchable logs to debug errors, monitor performance, and understand usage patterns.
Quickstart
Generate embeddings in a single request with your API key.
Simple endpoints
GET /v1/models to list models, POST /v1/embeddings to generate vectors.
Overflow strategies
fail, truncate, or chunk — control what happens when inputs are too long.
Custom dimensions
Reduce size with output_dimension for faster search and lower storage.
Usage info
Every response includes token usage for monitoring and cost estimation.
curl -X POST https://api.vectors.space/v1/embeddings \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{ "model": "embeddinggemma-300m", "provider": "llama", "input": "Your text here", "strategy": { "type": "truncate" } }'
Built by the team at
SQLite AI
We are building developer-first AI infrastructure. Our goal is to make the "data plane" of AI applications as reliable and predictable as a standard database.
Model Portability
Switch between OpenAI, Gemini, Voyage, or local models instantly. One unified API means you aren't tied to a single model provider's SDK.
Operational Stability
We focus on the boring parts of AI—uptime, latency, and caching—so you can focus on building your core product features.
Neutral Infrastructure
We provide the routing layer, not the models. We remain provider-neutral to ensure you always have access to the best performance.
