Periplo Cloud · early accessOpen-source coreRuns on your AWS

A senior engineering team, as a platform.

Periplo brings your lakehouse, pipelines, APIs and AI agents together on your own cloud. Everything it builds and runs stands on loom, our open-source foundation, so it ships with senior practices from day one. Polars by default; Spark when the volume asks for it.

Get early access
docker run --rm -p 127.0.0.1:8080:8080 ghcr.io/massivedatascope/periplo:latest

The command starts the open-source console on your machine. No sign-up.

Periplo console home: the lake grouped by layer, with tables, rows, size and last write
The open-source console, as it runs today, with sample data.

One platform · what's ready and what's next

● available · alpha

Data

See and query the Delta tables in your S3: schema, stats and history from the Delta log, and read-only SQL.

● early access · Cloud

Teams and security

Sign in with your identity provider, invite members, grant access, keep each client's data apart and audit who accessed what.

○ coming next

Pipelines

Follow and operate the ETL runs that write your tables, from the same console.

○ coming next

Build

Create a loom project from Periplo, deploy it to your AWS and watch it run in the same console.

○ coming next

AI

An agent that builds with loom's senior patterns, through typed changes you review.

The Periplo join builder: orders and customers joined on customer_id, checked as a many-to-one lookup, with the SQL it will run
Available: join without SQL, with the key proposed and the relation checked first. Sample data.
Preview of the Periplo ETL console: runs in the last 24 hours, ETLs that need attention and runs in progress
Coming next: the ETL console, in development. Sample data.

See the platform in detail →

Built on loom. Open source, all of it.

loom is the typed foundation under Periplo: one kernel for APIs, ETL, streaming, agents and jobs. Business rules, contracts and permissions are declared, compiled into a plan and checked before the first request. The same practices, whoever writes the code.

Use it on its own today. Apache-2.0.

pip install loom-kernel

Explore loom →

class CreateUserUseCase(UseCase[User, User]):
    rules = [
        Rule.check(F(CreateUser).email, via=_email_must_be_valid),
        Rule.forbid(lambda _, __, exists: exists,
                    message="email already exists")
            .from_params("email_exists"),
    ]

    async def execute(
        self,
        cmd: CreateUser = Input(),
        email_exists: bool = Exists(User, from_command="email",
                                    against="email"),
    ) -> User:
        return await self.main_repo.create(cmd)

Senior practices you can check.

We sell quality, so we show it. Every figure links to the public source where you can check it.

Databricks when you need it. Polars when you don't.

You declare a pipeline once and choose the engine in configuration. Most data jobs fit on one node and run on Polars for a fraction of the cost; Spark or Databricks take over when the working set outgrows it.

Small ETL every hour · 20 tables · ~2 GB · per month
Fargate ARM, 4 vCPU$11
Databricks Jobs$53
Daily ETL · ~50 GB · 30–45 min · per year
ECS r7g.4xlarge$175–262
Databricks Jobs$450–648

Open Source or Cloud

Periplo Open Source

The console and loom, for one engineer on one machine. You run it.

View on GitHub →

Periplo Cloud

Your team, your clients and their data, kept apart and audited. We run the control plane; your data stays in your S3. Prepaid, with a spending cap.

Get early access

Manifesto

Senior judgement doesn't scale. So we compile it.

Read the manifesto →

Who's behind Periplo

Periplo is made by MassiveDataScope, a software company based in Madrid, Spain. We publish what we build as open source: loom, the Periplo console, plywatch and loom-actions.

About MassiveDataScope →