Regulated document drafting
Batch-draft claims appeals, filings, or reconciliations with parallel agents — and hand over a verifiable audit bundle of exactly what ran. Read the use case →
Pangolin Scale
Submit a plan of agent tasks — they run in dependency order and fan out safely under resource locks. Get back reviewable patches and a verifiable audit trail — tamper-detecting by default, tamper-evident at the S3 Object Lock tier.
Who is it for?
Section titled “Who is it for?”Dev offload
Fan codebase maintenance out to sandboxed agents and get back reviewable patches — unattended, but never unaccountable. Read the use case →
Compliance evidence
Export a sealed evidence bundle per run; an auditor re-verifies it with one command — guarantee tiers stated honestly. Read the use case →
Data pipelines
The same provable engine for non-LLM batch jobs — typed handoffs, runtime fan-out, identical audit chain. Fully offline demo. Read the use case →
What do you want to do?
Section titled “What do you want to do?”Dispatch one agent
Wire PangolinClient into your app: register a capability, a subagent, and an env, then dispatch and read the result.
Start the dispatch tutorial →
Orchestrate a task plan
Run many agent tasks unattended with pangolin orch: tasks execute in dependency order, fan out in parallel under resource locks, and each drops a reviewable patch — with a verifiable audit bundle for the whole run.
Run your first offload →
Extend a provider
Plug in a new compute, storage, credential, or result-sink backend behind Pangolin Scale’s seams. Write a provider →
Evaluate / is it safe?
What Pangolin Scale is, how agents are sandboxed, the privilege boundary, and what the BSL license means for you. Read the architecture →
Ship Pangolin Scale in a product
Embedding Pangolin Scale in a product for your customers, or need enterprise compliance modules and support? Start a white-glove pilot. Commercial & pilots →