Skip to content

Introduction

bytetourist is a proxy & scraping network that you own end to end. Instead of renting bandwidth from a residential-proxy reseller, you run your own fleet of edge nodes — cloud VMs (and, later, mobile devices donating bandwidth) that make the final outbound HTTP call — and a central core that routes each request to the right node using one of ten selection strategies.

It is the layer that scraping APIs and “web unblockers” rent. bytetourist lets you operate it yourself: own the IPs, own the routing, own the margin.

  • Scraping & data teams who want residential/datacenter/mobile egress with real routing control, not a single rotating-proxy endpoint.
  • Platforms that resell proxy or scraping capacity and want to stop paying a ~30% bandwidth tax to upstream proxy vendors.
  • Anyone who needs geo-targeted, fingerprint-resistant outbound requests at scale, with per-tenant quotas and metering built in.

bytetourist is four small Go services around a few stateful dependencies:

Service Role
core The forward proxy. Authenticates, applies plan limits, matches a request to a node, and streams it out.
edge A node binary. Receives a request over a gRPC stream, makes the real call, streams the response back.
scaler Ingests usage, evaluates scaling rules, and provisions/recycles edge nodes across clouds.
server The dashboard & control-plane API (orgs, API keys, plans, usage).

Supporting infra: etcd (node membership), Redis (the broker — pub/sub and in-flight rate limiting), Kafka (the request-event log), ClickHouse (analytics), and Postgres (tenancy).

  • node — an edge node: the entity that makes the final outgoing call. It can be a cloud VM running the edge binary, or a mobile app donating bandwidth.
  • core — the central router clients send their proxy traffic to.
  • broker — the Redis instance used for pub/sub.
  • ip_type — the network class of a node’s egress IP: datacenter, residential, or mobile.

bytetourist sits one layer below scraping APIs like Firecrawl. Those tools turn a URL into clean data but buy their proxy bandwidth. bytetourist is the egress network itself — and a Scrape API (markdown + structured extraction on top of your own fleet) is on the roadmap.