Scalable Sandbox and Orchestration Infrastructure for Agents at https://nitter.gq/t.co/XyHdpcihXy
San Francisco, CA
Joined November 2023
- Tweets486
- Following70
- Followers1.7K
- Likes431
Exciting! Come and join us!
Giving a talk at Unfortunately Computers on October 29th about @tensorlake’s content-addressed block storage system and efficient snapshotting infrastructure for sandboxes.
If you love sandbox tech, please come say hi! unfortunatelycomputers.com/e…
Thanks @davidgu and @recallai for hosting.
1
2
963
Tensorlake is now a built-in sandbox provider in @scale_AI 's AgentEnv Framework. An open-source framework for building RL environments.
Check out the docs and the demo to see how the Tensorlake sandbox hosts a full environment for a PortSim task.
agentenvframework.com/docs/s…
Added @tensorlake as a sandbox provider to @scale_AI AgentEnv Framework. I tested it with a PortSim task:
An agent runs a port through a real windy week inside two Tensorlake sandboxes. Got a perfect score.
It was super fun to build and a great environment to play with.
Thanks for trying out Tensorlake and for the kind words! We'd love to see what you built. Let us know if we can help with anything!
I wanted to explore agent sandboxing through a small project, so I built a prototype using @pydantic and @tensorlake. It was a great experience with tensorlake.
Credit to @tensorlake for the excellent documentation, error responses and free tier to experiment.
2
402
Check out our docs on live resizing a Tensorlake sandbox!
docs.tensorlake.ai/sandboxes…
One of the things you can now do with Tensorlake's new VMM is grow your VM's disk size while it's running!
If your coding agent runs out of space when it's compiling a project, just resize and keep going.
1
253
We are working on supporting the Agent Sandbox Protocol (ASP) in Harbor!
We have an active RFC for an Agent Sandbox Protocol (ASP) that disentangles an agent's execution environment from its runtime. There are many advantages it provides, including preventing trajectory modification.
- increased security
- resource isolation
- no contamination
- separate configuration (e.g. no network in execution environment)
For this to work, we'd need agent developers to opt in and implement their tools in a sandbox-agnostic manner using the protocol. I personally think this is the direction the industry needs to move in, and would love the community's help in designing this protocol.
github.com/harbor-framework/…
Tensorlake Sandboxes are coming to Europe 🇪🇺
Reach out if you need EU data residency.
We are going to expand @tensorlake to Europe as we are seeing demand from users there! If you need EU residency for your sandboxes please reach out.
2
398
Tensorlake Sandboxes, now in @computesdk Actions 🚀
Run your CI/CD workloads on fast, elastic Tensorlake compute.
Great to see this integration live 🤝
@computesdk Actions 🤝 @tensorlake
Now you can run your actions directly on Tensorlake Sandboxes.
Tensorlake has worked tirelessly to provider elastic compute. Your CI/CD gets all of the benefits of their work now.
Coding agents need their own Git workspaces, not just sandboxes.
@sebuzdugan tested Tensorlake Git:
- fork a repository in ~160 ms,
- catch broken changes before merging,
- roll back a bad refactor in ~1 second. No running sandbox needed
A hands-on look at Git built for agents 👇
your agent made a mistake.
the file changed.
the database changed.
the tool already ran.
git can't help.
forking the whole environment can.
here's why the "undo button" for agents isn't actually a fork ↓
1
1
648
@tensorlake sandboxes now support streaming in Harbor 🎉
Just add:
`-e tensorlake --stream`
Check out @harborframework 's latest docs:
docs.harborframework.com/cor…
pr merged. congrats to @tensorlake on adding streaming support in harbor.
streaming now supports 4 providers: @daytonaio, @modal, @tensorlake and @Docker. If you want to merge yours, comment or DM me.
2
429
Love seeing Tensorlake benchmarked in the wild. The numbers speak for themselves 🔥 Thanks, @Offloop, for putting this together!
Just ran some benchmarks on @Offloop's sandbox providers and @tensorlake is crushing it.
Absolutely love what @diptanu and his team is doing!
1
4
743
How does traffic actually get from a public URL into an AI sandbox?
A new deep dive from @aiwithme0001 compares Tensorlake, E2B, Daytona, and Fly.io, including why Tensorlake moved part of its dataplane from L7 to L4.
Full comparison 👇
The full breakdown covers the mechanisms, trade-offs, and 4 questions to ask about any sandbox ingress hop.
Save or read on @towards_ai:
yadavdivy296.medium.com/ai-s…
In the latest @harborframework release, dynamic network policies now work with Tensorlake sandboxes.
Harbor lets an eval change network access across different phases of a trial:
setup → public network
agent → allowlist
verifier → no network
We updated the Tensorlake Harbor implementation to support this directly, so the same sandbox can move through each phase with a different network boundary without restarting or losing state.
More on how it works 👇
We now support running @cursor_ai Cloud Agents on Tensorlake sandboxes.
Check out the docs to get set up. docs.tensorlake.ai/sandboxes…
You can now run Cursor cloud agents on your infrastructure, including pools of machines that automatically scale with demand.
This lets you give agents access to internal services or specialized hardware, while the agent loop stays in Cursor.
2
410
@cursor_ai Cloud Agents now run on @tensorlake sandboxes.
Start 1 worker or fan out many. Each gets its own fast, isolated sandbox.
Enterprise teams can also run Tensorlake in their own VPC, so agents can reach private repos, APIs & internal services.
Tensorlake is back at #1 on ComputeSDK storage. 🥇
Storage lifecycle → 140 ms download, 598 Mbps
Snapshot & fork → 40 ms snapshot, 210 ms fork
Agents constantly save state, branch & run in parallel.
Fast storage primitives matter.
GPUs are now live on Tensorlake 🚀
It’s bringing our fast, stateful runtime to GPU workloads, with startup in 1–3 seconds.
If you’re running inference with vLLM or SGLang, we’d love to hear about your use case.
A small fleet of GPUs are in production @tensorlake and in use by our users!
The most challenging aspect of building this product was building a runtime which can be spun up in 1-3 seconds using a content addressable file system for storing images, CUDA and memory snapshots.
If you are doing running VLLM and SGLang for inference please reach out!
1
255
We added one SDK operation for this:
Python: Sandbox.get_or_create()
TypeScript: Sandbox.getOrCreate()
Pass it a stable name, such as an agent session ID. It returns a running sandbox by connecting to the existing one, resuming it when suspended, or creating it when it does not exist.
The SDK also handles concurrent create attempts and waits until the sandbox is ready. This removes that lifecycle code from agent harnesses and is particularly useful for long-running and background agents.
Available in tensorlake>=0.5.119.
One thing I noticed after building several sandbox integrations:
The harness/application still ends up managing parts of the sandbox lifecycle and the session-to-sandbox binding: reconnecting, resuming, handling restarts, and tracking which sandbox belongs to which session.
For agent workloads, I think the SDK should be closer to:
durable session identity → usable compute
1
3
514
Tensorlake's dynamic network policies let you change the network boundary on a running sandbox, which is especially useful for practical agent workflows.
- Coding agents need network access to clone repos and install dependencies.
- They probably shouldn’t keep unrestricted egress once they start executing model-written code.
setup → lock down → execute
Same environment. Different network policy.
Great hands-on deep dive from @sebuzdugan 👇
1
1
217
This is what powers the dynamic network policy support we added to @harborframework , allowing agents to switch network policies without restarting the running environment.
We shipped a comprehensive update to Tensorlake's egress traffic and firewall infrastructure for sandboxes.
Users can now update almost all aspects of traffic control while a sandbox is running. This lets a sandbox start up, download or configure tools, and then lock down before agents begin working in it.
Support for DNS-based policies improved as well:
1. Allow and deny based on DNS names (including wildcards).
2. Evaluate DNS-based flows per connection, so destination IP churn doesn't let traffic slip to denied endpoints.