Mirmir
mirmir.tech
Private Preview
A single place your data should be
Mirmir is a service for data analysis, embedding, and storage in a vector database. Built for developers and teams who want to instantly connect their data with AI.
SDK: TypeScript • Python • Rust • Go
Examples
use anyhow::Result;
use mirmir::Client;
#[tokio::main]
async fn main() -> Result<()> {
let api_key = std::env::var("MIRMIR_KEY")?;
let client = Client::new("https://api.mirmir.tech", api_key);
// Embed and upsert
let text = "hello world";
let vector = client.embed_text(text).await?;
client
.vectors()
.upsert("doc-42", &vector)
.metadata([("source", "docs/intro.md")])
.send()
.await?;
// Similarity query
let results = client
.query()
.similarity(&vector)
.top_k(5)
.send()
.await?;
for r in results {
println!("- {} ({:.2})", r.id, r.score);
}
Ok(())
}Analysis
Process data and extract features and metadata with built-in pipelines.
Embedding
Create vectors from text, images, and binary files. Choose model and precision.
Vector DB
Fast upserts, scalable indexes, and semantic ANN search.
Connectors and integrations
Plug & play: S3, GCS, PostgreSQL, BigQuery, Snowflake, Git, Notion. Webhooks and cron for continuous ingestion.
S3PostgresGitSnowflakeGCSBigQueryNotion
Automation and orchestration
Schedules, job queues, retries and idempotency. Webhooks for real-time data changes.
Security and compliance
Encryption at rest
Encryption in transit
Data residency
Key rotation
Audit and logs
WAF + DDoS
SOC2ISO 27001EU / GDPR
Access policies
- Project RBAC
- Scoped tokens
- File quarantine
- PII masking
Mirmir
mirmir.tech
Build an assistant on your data in 15 minutes
Private Preview launching soon. Leave your email to join the waitlist.
No spam. Email confirmation.
Plans
- HobbyFree
- Startup€49/m
- ScaleCustom