Use lightweight, programmable triggers to decide what data is important without overwhelming edge compute resources.
No more wondering whether data got uploaded. We track each piece of data through the entire upload and ingest process.
Develop triggers locally, backtest in the cloud, and push to edge devices over the air.
Any code that runs on edge devices is open source and typically runs inside a container.
Events like data upload triggers or data properties on ingest can be used to send you alerts.
Droplet comes with all the tools you need to collaborate with your team.
Quickly understand the impact of an issue with built in reporting.
Our powerful natural language search and filtering tools allow you to separate the data that matters from everything else.
Most data isn't useful. Set policies to control data retention by age, tags, and inclusion in datasets.
Easy-to-configure RBAC allows you to configure what data can be accessed by individuals, teams, and automated systems.
Real world AI needs to be auditable, which means data needs to be too. Trace samples back to the source.
Automatically triage new recordings by applying tags based on the content of the recording and other metadata.
Automate the process of curating, cleaning, and adding new examples to datasets.
Chain multiple transformations to build PII redaction pipelines, generate embeddings, and produce arbitrary derived data.
Automatically ingest and convert between rosbags, mcaps, video, text, and other formats.
Avoid downloading entire recordings. Our data API lets you access only what you need, filtering by time range, data fields, and doing format conversions on the fly.
Our product is built to be pluggable and hackable. We provide REST APIs to control Droplet from your system, and webhooks if you want Droplet to talk back.
hi@dropletdata.com
