Cloud Agents Platform

Agents that keep working when you close the tab.

LeelE runs long-lived AI agents on infrastructure you own. Work checkpoints as it happens, memory persists across sessions, destructive actions wait for your approval, and every token is accounted for.

research-synthesis12.4k tok · $0.031
  1. plancheckpoint 1
  2. search:webcheckpoint 2
  3. read:sourcescheckpoint 3
  4. redeployprocess replaced
  5. resumefrom checkpoint 3
  6. synthesiserunning · 4.2s
  7. publish:reportawaiting approval

A real run: three checkpoints, a platform redeploy that replaced the process mid-flight, and a resume from the last checkpoint — not a restart.

What it does

Everything a long-running agent needs to be trusted with real work.

Durable runs

Work checkpoints as it happens, so a run survives a deploy, a crash or a wait for approval — and resumes from where it stopped rather than starting over.

Memory that lasts

Conversations keep their history, recall what matters from earlier, and compact the rest, so context stays useful well past the point a window would overflow.

Approvals before actions

Destructive tools stop and ask. Confirmations are cryptographically signed, so an approval cannot be forged, and every decision is recorded.

Costs you can see

Every run is attributed by user, model and token type — including cached tokens — so spend is measured rather than estimated.

Parallel subagents

Focused child agents run alongside the main thread with constrained tools, and report back without flooding the conversation.

Yours to run

Self-hosted on infrastructure you control, with your own model provider. No data leaves your account by default.

How it works

Two execution planes, because a chat reply and an overnight job are not the same problem.

Most platforms pick one and make the other awkward. Interactive turns take the direct path; durable work checkpoints so it can be interrupted without loss.

Interactive

Answers while you wait

A chat turn runs in-request and streams straight back. No queue, no scheduler, no indirection — the fastest path is the direct one.

  • Streams token by token
  • Resumes if the connection drops
  • Bounded by an explicit time budget
Durable

Work that outlives the request

Long and scheduled work checkpoints per turn. Between turns nothing is running, so a job can wait hours for a subtask or days for an approval and still be one run.

  • Survives deploys and restarts
  • Retries a turn, never the whole run
  • Suspends free while awaiting approval

Access

Built in the open, deliberately.

LeelE ships in small, verified increments. Access is granted by invitation while the platform is early — not to manufacture scarcity, but because every account is one we can actually support.

LeelE is invite-only. Requests are reviewed by hand.