splime.io

What It Is

splime positions itself as a “Private Python execution network.” It is not a general-purpose cloud functions platform. Instead, it packages a team’s existing trusted Python functions or SPL pipelines into immutable versioned objects, syncs them through a central registry, and lets a daemon agent on private machines connect outbound, claim tasks, and execute them. The goal is to make internal code scattered across legacy repositories and notebooks discoverable, reusable, and auditable.

Core Capabilities

Functionally, splime supports serializing functions into SPL/YAML bundles, including inputs, outputs, dependency manifests, and environment information. At runtime, users can specify a target_machine, build an isolated venv on the machine where the data resides, and execute the job there. The console provides visibility into machines, object versions, run status, timelines, logs, results, and artifacts. Its security model emphasizes scoped tokens, delegated machine access, immutable versions, and auditable run history. The worker agent does not need to expose itself to the public internet; it only needs an outbound connection to the server.

Language, API, and Ecosystem

The available materials only explicitly mention Python support, making it suitable for data processing, model scoring, validation, report generation, automation scripts, and similar use cases. For developer experience, splime provides Python SDK examples: after initializing SPLClient, users can register_env, publish, and call, with a relatively small API surface. In terms of ecosystem, the disclosed components currently mainly include notebooks, services, the daemon agent, the console, and the registry. There is no visible documentation yet for CI/CD, identity provider, cloud platform, or data stack integrations.

Pricing and Deployment

splime is currently in private beta and requires an access request. Public plans, pricing, free quotas, and payment methods have not been disclosed. Architecturally, it appears to use a hosted control plane for coordination while user-owned private workers perform execution. The page does not confirm whether the server side can be self-hosted, so teams with strict compliance requirements or fully offline deployment needs will need further discussion with the vendor.

Pros, Cons, and Who It’s For

Its main strengths are lightweight reuse of internal Python code without immediately adopting heavier orchestration systems such as Airflow or Dagster, while also addressing private network boundaries, version reproducibility, and artifact management. Limitations include uncertain maturity during the beta phase, unclear open-source status, SLA, compliance, security details, and pricing, as well as currently narrow language coverage. It is best suited for data/ML, platform engineering, and automation teams looking to run a pilot.

Access from China

The page does not provide information about China-region nodes, ICP filing, payment options, or network connectivity, so access from China can only be rated as unknown. Domestic teams looking for alternatives may evaluate Prefect, Dagster, Airflow, Ray, Celery, Runhouse, and similar tools, combined with deployment in their own network environment.