AI Potluck
Infrastructure / Deployment

Tensorlake Sandbox

Tensorlake

Serverless Firecracker-microVM sandbox runtime with durable-execution orchestration for agentic apps, offering snapshot/suspend/resume and scale to thousands of sandboxes. The runtime and SDK are Apache-2.0 with a bring-your-own-cloud self-host path (run sandboxes inside your own AWS/GCP/Azure account); the managed scheduler is the hosted layer. Backed by Redpoint and Amplify Partners; the company repositioned from document ingestion to a sandbox-native cloud. ~959 stars on GitHub.

Tensorlake sandbox-native cloud for AI agents; repo Apache-2.0 (Python + Rust), CLI v0.5.65 live on GitHub July 2026. Firecracker microVMs with BYOC self-host; managed scheduler (Lattice) proprietary.

Openness

4 medium confidence
4.0
license
Apache-2.0(OSI)
source
public(runtime+SDK, BYOC self-host in your own AWS/GCP/Azure)
managed-tier
Tensorlake hosted cloud + proprietary Lattice scheduler
core-gated
gated

Apache-2.0 runtime and SDK with a genuine bring-your-own-cloud self-host path, plus a proprietary managed scheduler, so open-core on par with the permissive-Apache peers (E2B). The open part is permissive and self-hostable; the managed scheduler is the closed value-add.

Adoption

2 low confidence
2.0

No published usage or download figure for the runtime; ~959 GitHub stars is the available signal, alongside Redpoint and Amplify backing and a free tier. stars_fallback caps the level at 3; placed at 2 on a low-thousands star base absent any volume signal. Low confidence.

Capability

4 medium confidence
4.0

Firecracker microVM isolation plus durable-execution orchestration and snapshot/resume make for a rich agent-runtime feature set, level with the microVM peers. Independent ComputeSDK TTI benchmarks measure ~460ms median cold-start (12th of 19 sandbox providers), mid-pack and consistent with the sub-second claim.

  • https://www.tensorlake.ai/ recorded 2026-07-09

    Firecracker microVM sandboxes, snapshot/suspend/resume, durable execution, scale to thousands

  • https://www.computesdk.com/benchmarks/sandboxes/ recorded 2026-07-09

    ComputeSDK TTI leaderboard (independent, 100 iters/day, 2026-07-09 run): median time-to-interactive ~460ms sequential / ~960ms burst @100 concurrent; 12th of 19 providers

Unchanged since 2026-07-30 (last edited, not re-checked)