Tensorlake Documentation

Use read-only mounts when a sandbox needs files but should not write to them. There are two read-only shapes:

Following: tracks the file system’s current state and refreshes as replicated autosaves or permanent snapshots land.
Pinned: resolves a specific snapshot or retained autosave once and never changes.

Following Read-only Mount

A following mount is best for shared skills, prompts, docs, configs, and dependencies.

CLI
Python
TypeScript

$ tl fs mount agent-assets /skills --ro
Mounted agent-assets at /skills (read-only, follows the filesystem)

from tensorlake.filesystem import FilesystemClient

client = FilesystemClient()
mount = client.mount("agent-assets", "/skills", readonly=True)

import { FilesystemClient } from "tensorlake";

const client = new FilesystemClient();
const mount = await client.mount("agent-assets", "/skills", true);

When the shared timeline advances, Tensorlake refreshes only the paths that changed. Unchanged files keep their warm cache.

Pinned Read-only Mount

A pinned mount is best for reproducible builds, evals, benchmarks, and released assets.

$ tl fs mount agent-assets:9f2a1c8e4d6b1a0f3c7e9d2b8a4f6c1e0d3b7a99fedcba987654321001234567 /release --ro

Use a permanent snapshot from tl fs history when an input must remain available indefinitely. A recent autosave can also be mounted, but it is an ephemeral recovery point and may age out of retention. For a complete asset distribution workflow, see Distribute Files to Agents.

Choosing Between Them

Need	Use
Every run must see the same files	Pinned read-only mount
Many sandboxes should receive updates without image rebuilds	Following read-only mount
An agent needs to write files	Writable mount

Overview

Guides

Reference

Use Cases

Read-only Mounts

Following Read-only Mount

Pinned Read-only Mount

Choosing Between Them

​Following Read-only Mount

​Pinned Read-only Mount

​Choosing Between Them

Following Read-only Mount

Pinned Read-only Mount

Choosing Between Them