tenki-mcp
This server lets an agent manage Luxor/Tenki sandbox microVMs and perform code execution, file, Git, networking, storage, template, SSH, and workspace operations.
Check auth status and identify the current credential and workspaces.
Run one-shot shell, Python, or JavaScript in a disposable sandbox.
Create, inspect, list, pause, resume, extend, update, and terminate persistent sandboxes, including bulk termination.
Execute commands in existing sandboxes and read, write, list, stat, create directories, move, and delete files.
Run supported Git operations: clone, checkout, diff, and log.
Expose ports and create, list, get, delete, touch, bind, unbind, and resolve preview URLs/tokens.
Create, inspect, list, update, and delete snapshots, and get snapshot download URLs.
Create, inspect, list, rename, resize, attach, detach, and delete persistent volumes.
Create, inspect, list, update, delete, build, cancel, and inspect template builds.
Read and update workspace usage, quotas, settings, and retention policies.
Get signed upload/download URLs for binary artifacts.
Configure SSH public keys, issue SSH certificates, and list SSH gateways.
Provides Git integration inside Tenki sandboxes, enabling agents to clone repositories and run checkout, diff, and log operations.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@tenki-mcp Run this Python in a fresh sandbox and tell me what it prints."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Luxor MCP: @luxorai/mcp
A Model Context Protocol server for Luxor Sandbox, part of Luxor Agent Cloud by Luxor AI. Give any agent (Claude, Codex, Cursor) a disposable microVM it can create, run code in, read and write files, run git, and expose to the web. Sandboxes boot in ~2 seconds and are billed per second.
The agent writes code and Luxor runs it in isolation. (Luxor Code Review and Luxor's managed CI runners are separate products; this server currently exposes the Sandbox.)
The old @tenkicloud/mcp name keeps working and publishes the same code.
"Run this Python in a fresh sandbox and tell me what it prints."
│
▼ tenki_run_code
boots a microVM → runs it → returns stdout → tears it downQuickstart
export LUXOR_API_KEY=lx_your_key_here
npx -y @luxorai/mcp # speaks MCP over stdioNothing to clone or build. The package installs the command luxor-mcp (and tenki-mcp, the same program).
Use it in Claude Code
Connect to Luxor's hosted MCP service and sign in with your Luxor account:
claude mcp add --transport http luxor https://mcp.luxor.ai/mcp
claude mcp login luxorClaude Code's mcp add command stores the server configuration; mcp login performs the one-time browser authorization and workspace selection.
Alternatively, run the MCP server locally with an API key:
claude mcp add luxor --env LUXOR_API_KEY=lx_your_key_here -- npx -y @luxorai/mcpOr install it as a Claude Code plugin
The repo doubles as a plugin marketplace. This prompts for your API key on install and stores it in your OS keychain, so there is no env var to manage:
/plugin marketplace add LuxorLabs/tenki-mcp
/plugin install luxor@luxorIf you added this marketplace before it was named luxor, your installs keep working as luxor@tenki and tenki@tenki. To switch to the new name, run /plugin marketplace remove tenki (this uninstalls its plugins and their saved API key), then add the marketplace again.
The luxor plugin names its tools luxor_*; the legacy tenki plugin keeps tenki_*.
Use it in Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"luxor": {
"command": "npx",
"args": ["-y", "@luxorai/mcp"],
"env": { "LUXOR_API_KEY": "lx_your_key_here" }
}
}
}Use it in Cursor
Add the same block to .cursor/mcp.json (project) or ~/.cursor/mcp.json (global):
{
"mcpServers": {
"luxor": {
"command": "npx",
"args": ["-y", "@luxorai/mcp"],
"env": { "LUXOR_API_KEY": "lx_your_key_here" }
}
}
}Use it in Codex
The OpenAI Codex CLI reads MCP servers from ~/.codex/config.toml. Add:
[mcp_servers.luxor]
command = "npx"
args = ["-y", "@luxorai/mcp"]
env = { LUXOR_API_KEY = "lx_your_key_here" }Confirm it's working
After adding the server, start a fresh session so the client loads it, then ask your agent one of:
"Check Luxor auth status.": confirms your credential was picked up.
"Run
print(2+2)in a fresh Luxor sandbox.": a full boot → run → teardown round-trip.
If the tools don't show up, restart the client so it re-reads its config.
No key yet? Start the server without one and ask the agent to check tenki_auth_status; it reports what to set and where. See Auth.
Run it from a clone instead
For development, or to run an unreleased change:
git clone https://github.com/LuxorLabs/tenki-mcp.git && cd tenki-mcp
npm install && npm run build
LUXOR_API_KEY=lx_your_key_here node dist/index.jsSubstitute node /absolute/path/to/tenki-mcp/dist/index.js for the npx command in any of the configs above. See CONTRIBUTING.md for the full development loop.
Environment variables
Variable | Default | Purpose |
| - | API key ( |
| - | Session token ( |
|
| Control-plane base URL ( |
| off |
|
| - | Comma-separated tool names to skip registering. |
| off |
|
|
|
|
|
|
|
|
| HTTP transport port. |
|
| HTTP bind host; non-loopback requires |
| - | Bearer token for the HTTP endpoint; optional on loopback, required on a non-loopback host. |
| - | Public base URL for an OAuth-protected hosted server. |
| - | OAuth authorization-server issuer. Enables delegated OAuth HTTP mode. |
|
| RFC 8707 resource identifier accepted in access-token audiences. A comma-separated list serves one resource per domain; metadata and the 401 |
|
| Required delegated scope. |
|
|
|
| - | Internal Luxor Identity service endpoint used to exchange OAuth access tokens. |
| - | Service credential for the private Identity token-exchange RPC. |
Every LUXOR_* variable above is still accepted under its TENKI_* name (for example TENKI_API_KEY, TENKI_MCP_TRANSPORT). All LUXOR_* credentials are read before any TENKI_* one. Keys start with lx_; tk_… keys are still accepted. LUXOR_API_KEY is only used when it starts with lx_ or tk_; any other value is ignored.
Related MCP server: MileHost MCP Server
Tools
82 tools: all 76 tool-worthy public unary API methods (enforced by a CI parity audit, which also fails if a tool calls a method the API no longer has; GitOperation is exposed as one tool per operation), two workflow helpers, and tenki_auth_status. Tool names use the tenki_ prefix by default; set LUXOR_MCP_TOOL_PREFIX=luxor (as the luxor plugin does) to name them luxor_*. A few admin surfaces (published images, saved Tailscale connections) and implementation-only control-plane methods are intentionally excluded. Grouped by domain:
Domain | Tools |
Auth |
|
Identity |
|
Run |
|
Sandboxes |
|
Session admin |
|
Exec |
|
Files |
|
Git |
|
Ports & previews | expose · list-exposed · unexpose · create-preview-url · open-preview (viewer-token-gated, not public; port 7681 is the web terminal) · list/get/delete-preview-url · touch-preview · bind/unbind-preview-url · resolve-preview-token |
Artifacts (binary transfer) |
|
SSH |
|
Snapshots | create · get · list · list-session · list-workspace · list-dangling · update · delete · get-download-url |
Volumes | create · get · list · update · delete · resize · attach · detach |
Egress policies | create · get · list · update · delete (named workspace allowlists that |
Templates | create · get · list · update · delete · build · cancel-build · get-build · list-active-builds · list-builds · delete-build |
Workspace |
|
Full per-release breakdown in CHANGELOG.md.
Deprecated upstream, still served: tenki_expose_port / tenki_unexpose_port / tenki_list_exposed_ports (the API prefers the preview-URL tools), tenki_list_workspace_sandboxes / tenki_list_workspace_snapshots (prefer the credential-scoped lists), and the ExecuteCommand RPC behind tenki_exec / tenki_run_code / tenki_move_path (successor is the data-plane Run stream, which needs an HTTP/2 transport this server does not have yet). The parity audit prints these on every run.
Auth
Set one of LUXOR_API_KEY or LUXOR_AUTH_TOKEN (TENKI_API_KEY / TENKI_AUTH_TOKEN are still accepted). When both are set, the auth token wins. The header is chosen by token prefix: lx_… or tk_… → Authorization: Bearer, ory_st_… → X-Session-Token, otherwise a session cookie. Override the endpoint with LUXOR_API_ENDPOINT (default https://api.luxor.ai, the same API as https://api.tenki.cloud).
Without a credential the server still starts, registering only tenki_auth_status, so instead of an MCP client reporting an opaque "server failed to start", the agent can call that tool and get told what to set. Ask it "check Luxor auth status" any time other tools return auth errors: it reports the credential kind (API key vs session token), the endpoint, and whether a live identity probe succeeded, never the token itself. It reports status only; get a credential from the Luxor dashboard.
Host it over HTTP (v2.0-beta)
Besides stdio, the server speaks Streamable HTTP so it can be hosted for remote MCP clients:
LUXOR_MCP_TRANSPORT=http PORT=3000 LUXOR_API_KEY=… npx -y @luxorai/mcp
# → tenki-mcp running on http://127.0.0.1:3000/mcp (Streamable HTTP) [loopback only, no auth]For a single-user deployment, HTTP mode can hold one shared LUXOR_API_KEY and therefore exposes a powerful capability. By default it:
binds to loopback (
127.0.0.1) only: setLUXOR_MCP_HTTP_HOST=0.0.0.0to expose it, but thenit requires a bearer token: set
LUXOR_MCP_HTTP_TOKENand sendAuthorization: Bearer <token>. It refuses to start on a non-loopback host without one.DNS-rebinding protection is on (Host-header allowlist), so a malicious web page can't drive your local server.
# expose to a network safely:
LUXOR_MCP_TRANSPORT=http LUXOR_MCP_HTTP_HOST=0.0.0.0 PORT=3000 \
LUXOR_MCP_HTTP_TOKEN=$(openssl rand -hex 32) LUXOR_API_KEY=… npx -y @luxorai/mcpPoint an HTTP-capable MCP client at /mcp. Static-key mode uses one shared LUXOR_API_KEY for all sessions. Verified end-to-end (test/http-transport.test.mjs: auth gate, DNS-rebinding rejection, connect → tools/list → tool call over HTTP).
Hosted multi-tenant deployments instead use Luxor Identity's OAuth facade. The MCP server publishes RFC 9728 protected-resource metadata and exchanges each caller's access token through the private Identity service. Identity owns Hydra, validates the requested audience and scope, and issues a short-lived API delegation bound to the user, client, and workspace selected on the Luxor consent page. tenki-mcp has no Hydra Admin access and does not hold the delegation-signing secret.
How it works
Luxor's API is ConnectRPC: JSON over HTTP/1.1, not REST. Every control-plane call is POST https://api.luxor.ai/tenki.sandbox.v1.SandboxService/{Method} with a lowerCamelCase JSON body. Per-session file I/O runs on a separate data-plane endpoint returned at create time, authenticated with a short-lived session certificate. This server owns both transports so the tools stay one-liners.
Command output: tenki_exec and tenki_run_code capture stdout/stderr by redirecting to files (sh -c '… > out 2> err') and reading them back over the data plane, so you get the output inline through a plain HTTP client.
The wire details are ported from the live-verified n8n community node.
Security
This server holds a Luxor API key and can run code + spend credits, so treat it as a capability. Full model + CSA MCP Server Top-10 mapping in SECURITY.md. Quick controls:
Least privilege: every tool carries a title and MCP annotations:
readOnlyHinton reads, anddestructiveHinton every tool that can overwrite, modify, interrupt or delete existing state (only tools that create, resume, extend, attach or keep alive without changing existing state are non-destructive). RunLUXOR_MCP_READONLY=1for an inspection-only server (read tools only), orLUXOR_MCP_DISABLED_TOOLS=tenki_run_code,…to drop specific tools.HTTP transport is loopback-only by default and requires a bearer token to expose to a network (see Host it over HTTP).
Audit:
LUXOR_MCP_AUDIT=1logs each tool call's name to stderr.Untrusted output:
tenki_run_code/tenki_exec/tenki_read_filereturn output from untrusted code; clients should treat tool results as data, not instructions.
Report vulnerabilities via a private security advisory, not a public issue.
Related
Luxor Sandbox, the platform: https://luxor.ai
n8n-nodes-tenki, Luxor Sandbox as an n8n node: https://github.com/opencolin/n8n-nodes-tenki
License
MIT
Available Tools
71 toolstenki_attach_volumeA
Mount a volume into a running sandbox at an absolute path. Set read_only to mount without write access.
Inputs: volume_id (required), mount_path (required), session_id (required), read_only.
tenki_auth_statusARead-onlyIdempotent
Report whether the server has a usable Tenki credential, which kind (API key vs OAuth session token), and the endpoint it targets — verified with a live identity probe. Call this first when other tools fail with auth errors, or when this is the only tool available (which means no credential is configured). Reports status only; it does not log in and never returns the token.
Inputs: no named parameters. Returns: error, detail, source, endpoint, +4 more
tenki_bind_preview_urlA
Bind a named preview URL to a sandbox session and port (advanced routing).
Inputs: port (required), session_id (required), preview_url_id (required).
tenki_build_templateA
Trigger a build for a template, producing a bootable image. Returns the created build — poll it with tenki_get_template_build until READY; the ready build's imageDigestRef is what tenki_create_sandbox's image arg takes.
Inputs: template_id (required), build_env, image_name, build_secrets, +1 more.
tenki_cancel_template_buildA
Cancel an in-progress template build by its build ID.
Inputs: build_id (required).
tenki_create_preview_urlA
Create a shareable public preview URL for a port in a sandbox. The sandbox must have inbound networking enabled (create it with allow_inbound).
Inputs: port (required), slug (required), session_id (required), expires_at.
tenki_create_sandboxA
Create a persistent sandbox microVM, optionally from a snapshot or template image. Returns the session (id, state) and its data-plane endpoint. Boots in ~2s. Use tenki_exec / tenki_read_file / tenki_write_file against the returned session_id.
Inputs: env, name, tags, image, +11 more.
tenki_create_snapshotA
Capture a running sandbox's disk and memory as a reusable snapshot (attached volumes are NOT captured); boot a new sandbox from it later with tenki_create_sandbox + snapshot_id.
Inputs: session_id (required), name, expires_at, store_raw_image.
tenki_create_templateA
Create a custom-image template (a reusable sandbox-image spec: base image + setup script + default resources). Build it into a bootable image later with tenki_build_template. NOTE: only a TYPED template (created with builder_spec, no legacy fields) can build a named, publishable image (image_name) that tenki_create_sandbox boots via its image arg.
Inputs: name (required), tags, env_vars, cpu_cores, +9 more.
tenki_create_volumeA
Create a workspace-scoped persistent volume — durable block storage that survives sandbox teardown. Defaults the workspace to the API key's first; override with workspace_id.
Inputs: name (required), size_bytes (required), workspace_id.
tenki_delete_preview_urlADestructive
Delete a preview URL by id, taking it permanently offline.
Inputs: preview_url_id (required).
tenki_delete_snapshotADestructive
Permanently delete a snapshot by ID.
Inputs: snapshot_id (required).
tenki_delete_templateADestructive
Delete a template by ID. Pass force to delete even when builds or dependents exist.
Inputs: template_id (required), force.
tenki_delete_volumeADestructive
Permanently delete a volume and destroy its data. Fails with VolumeInUse if the volume is still attached to a session — detach it first.
Inputs: volume_id (required).
tenki_detach_volumeBDestructive
Unmount a volume from a sandbox session.
Inputs: volume_id (required), session_id (required).
tenki_execA
Run a command in an existing sandbox and return stdout, stderr, and exit code inline. Streams over max_output_bytes (default 64KB) come back as a head+tail preview with the full output retained at stdoutPath/stderrPath in the sandbox.
Inputs: command (required), session_id (required), cwd, env, +3 more. Returns: ok, args, stderr, stdout, +7 more
tenki_expose_portA
Expose a port from a sandbox and get a public preview URL. Useful when an agent starts a web server it wants to show.
Inputs: port (required), session_id (required), slug.
tenki_extend_sandboxA
Extend a running sandbox's wall-clock lifetime by N seconds so it isn't auto-terminated at its max-duration cap.
Inputs: session_id (required), additional_duration_seconds (required).
tenki_get_download_urlARead-onlyIdempotent
Get a short-lived signed URL to download (HTTP GET) a command artifact from a sandbox by its artifact id (e.g. a command's stdout/stderr artifact). Note: the API supports download-by-artifact-id only, not download-by-path.
Inputs: session_id (required), artifact_id (required).
tenki_get_preview_urlARead-onlyIdempotent
Fetch a specific preview URL's details by id.
Inputs: preview_url_id (required).
tenki_get_sandboxBRead-onlyIdempotent
Fetch a sandbox's current state and metadata.
Inputs: session_id (required).
tenki_get_snapshotARead-onlyIdempotent
Fetch one snapshot's status and metadata by ID.
Inputs: snapshot_id (required).
tenki_get_snapshot_download_urlARead-onlyIdempotent
Get a short-lived, pre-signed URL to download a snapshot's raw disk image (requires the snapshot to have been created with store_raw_image).
Inputs: snapshot_id (required).
tenki_get_snapshot_retention_settingsARead-onlyIdempotent
Get the workspace's pause- and snapshot-retention periods. DEPRECATED upstream: this RPC is marked deprecated in the API — tenki_get_workspace_settings returns the same retention fields alongside the quotas. An empty response means no retention override, i.e. kept indefinitely.
Inputs: workspace_id.
tenki_get_templateARead-onlyIdempotent
Retrieve one template by ID.
Inputs: template_id (required).
tenki_get_template_buildARead-onlyIdempotent
Retrieve one template build by its build ID (state, progress, and result image). A READY build's imageDigestRef is the reference tenki_create_sandbox's image arg takes.
Inputs: build_id (required).
tenki_get_upload_urlA
Get a short-lived signed URL to upload (HTTP PUT) a binary file to a path inside a sandbox. Use for non-text payloads too large or binary for tenki_write_file.
Inputs: path (required), session_id (required), content_type.
tenki_get_volumeARead-onlyIdempotent
Fetch a single volume's metadata and current state by its id.
Inputs: volume_id (required).
tenki_get_workspace_settingsARead-onlyIdempotent
Read a workspace's sandbox quotas and retention policy: max snapshots/templates/volumes/total bytes, max concurrent and sticky sessions, max preview URLs, and the pause/snapshot retention periods. These are workspace limits — there are no per-session defaults (idle timeout and max duration are set per sandbox at creation).
Inputs: workspace_id.
tenki_get_workspace_usageARead-onlyIdempotent
Get per-second sandbox billing and usage figures for a workspace — use this for cost visibility across all of the workspace's sandboxes.
Inputs: workspace_id.
tenki_gitA
Run a git operation in a sandbox. Only clone, checkout, diff, and log are supported by the API — for any other git command (status, add, commit, push, ...) use tenki_exec with git .... clone arg keys: repo (required, the URL), branch?, depth?, directory?. CAVEAT (live-verified): checkout/diff/log run in the session's working directory (/home/tenki), which is not a repository and has no directory arg — so on a repo cloned into a subdirectory they fail with 'not a git repository'. Use tenki_exec with git -C <directory> ... instead (e.g. git -C /home/tenki/hw log -n 2). For reference, their arg keys are — checkout: ref (required), create? ('true' = -b); diff: range? or base?+head?, path?; log: max_count?, range?, path?.
Inputs: operation (required), session_id (required), args.
tenki_issue_ssh_certA
Issue a short-lived SSH certificate for a public key, authorizing SSH access to a sandbox via the SSH gateway.
Inputs: public_key (required), session_id (required).
tenki_list_active_template_buildsARead-onlyIdempotent
List the currently active (in-progress) builds for a given template.
Inputs: template_id (required).
tenki_list_dangling_snapshotsARead-onlyIdempotent
List dangling snapshots — those whose source session no longer exists — for the workspace, useful for cleanup.
Inputs: page_size, page_token.
tenki_list_exposed_portsARead-onlyIdempotent
List the ports currently exposed from a sandbox.
Inputs: session_id (required).
tenki_list_filesARead-onlyIdempotent
List a directory in a sandbox, including dotfiles (.git, .env, .gitignore) by default — set include_hidden false to omit them.
Inputs: path (required), session_id (required), include_hidden.
tenki_list_preview_urlsARead-onlyIdempotent
List the workspace's preview URLs, newest page first. Pass session_id to keep only the ones bound to that sandbox (filtered here, not server-side — so it applies to the page you fetched; raise page_size or follow next_page_token to widen it). Results are paginated: a nextPageToken in the response means more pages exist.
Inputs: page_size, page_token, session_id, workspace_id.
tenki_list_sandboxesCRead-onlyIdempotent
List sandboxes for the workspace.
Inputs: page_size, page_token, include_terminated.
tenki_list_session_snapshotsARead-onlyIdempotent
List the snapshots captured from a specific sandbox session.
Inputs: session_id (required), page_size, page_token.
tenki_list_snapshotsBRead-onlyIdempotent
List the saved snapshots for the workspace (owner inferred from the API key).
Inputs: page_size, page_token.
tenki_list_ssh_gatewaysARead-onlyIdempotent
List the currently active SSH gateways for the workspace.
Inputs: workspace_id.
tenki_list_templatesARead-onlyIdempotent
List templates for the workspace, optionally filtered by tags.
Inputs: tags, page_size, page_token, workspace_id.
tenki_list_volumesARead-onlyIdempotent
List persistent volumes in a workspace (defaults to the key's first workspace). Supports pagination.
Inputs: page_size, page_token, workspace_id.
tenki_list_workspace_sandboxesARead-onlyIdempotent
List every sandbox belonging to a specific workspace (defaults to the API key's workspace) — useful for spotting leaked, still-billing sandboxes across the workspace.
Inputs: page_size, page_token, workspace_id, include_terminated.
tenki_list_workspace_snapshotsARead-onlyIdempotent
List all snapshots in a workspace (defaults to the key's first workspace). Supports pagination.
Inputs: page_size, page_token, workspace_id.
tenki_make_dirA
Create a directory in a sandbox. Set recursive to also create any missing parent directories (mkdir -p).
Inputs: path (required), session_id (required), recursive.
tenki_move_pathB
Move or rename a file or directory within a sandbox. Both paths are under /home/tenki.
Inputs: to (required), from (required), session_id (required).
tenki_open_previewA
Open a viewer-token-gated (AUTHENTICATED-mode) preview. USE tenki_expose_port OR tenki_create_preview_url INSTEAD for an ordinary web server: for any port other than the web terminal (7681) the API deliberately returns a non-regional fallback host that currently has no edge route, so the URL 404s (live-verified). The returned viewerToken does resolve via tenki_resolve_preview_token; only the URL is unreachable. Requires allow_inbound.
Inputs: port (required), session_id (required), expires_at.
tenki_pause_sandboxA
Pause a sandbox (snapshot + suspend) so it can be resumed later.
Inputs: session_id (required).
tenki_read_fileARead-onlyIdempotent
Read a UTF-8 text file from a sandbox (paths under /home/tenki).
Inputs: path (required), session_id (required).
tenki_remove_pathADestructive
Delete a file or directory in a sandbox. Set recursive to remove a non-empty directory and its contents (rm -r).
Inputs: path (required), session_id (required), recursive.
tenki_report_sandbox_activityA
Report client-side activity on a sandbox to reset its idle timer and keep it from being reaped as idle (a keep-alive heartbeat).
Inputs: session_id (required).
tenki_resize_volumeA
Grow a volume to a new size in bytes (1 MiB … 100 GiB). Volumes can grow but not shrink.
Inputs: volume_id (required), size_bytes (required).
tenki_resolve_preview_tokenARead-onlyIdempotent
Resolve a preview token to the sandbox/port it points at (advanced).
Inputs: token (required).
tenki_resume_sandboxA
Resume a previously paused sandbox.
Inputs: session_id (required).
tenki_run_codeA
Boot a throwaway microVM, run a snippet (shell/python/javascript), return its stdout/stderr/exit code, and tear the sandbox down. Cost-guarded and self-terminating. Use this for one-shot execution when you don't need a persistent sandbox. Output over ~64KB per stream is truncated head+tail — and the sandbox is gone, so for large output use tenki_create_sandbox + tenki_exec and page through the retained file.
Inputs: code (required), language (required), env, timeout_seconds.
tenki_stat_pathARead-onlyIdempotent
Get metadata (size, mode, type, timestamps) for a file or directory in a sandbox. Use to check whether a path exists or inspect it before reading/removing.
Inputs: path (required), session_id (required).
tenki_terminate_sandboxADestructive
Terminate (destroy) a sandbox. The microVM and its filesystem are gone after this.
Inputs: session_id (required).
tenki_terminate_sandboxesADestructive
Terminate MULTIPLE sandboxes in one call (bulk). IRREVERSIBLE — every listed sandbox and its filesystem is destroyed. Use tenki_terminate_sandbox for a single one.
Inputs: session_ids (required).
tenki_touch_previewA
Refresh (keep-alive) a live preview by its preview token so it isn't torn down as idle.
Inputs: preview_token (required).
tenki_unbind_preview_urlADestructive
Unbind a named preview URL from its current session/port (advanced routing).
Inputs: preview_url_id (required).
tenki_unexpose_portADestructive
Remove an inbound port exposure from a sandbox, taking its public URL/preview offline. Use this to un-publish a port previously exposed with tenki_expose_port.
Inputs: port (required), session_id (required).
tenki_update_sandboxA
Update mutable fields on an existing sandbox — its name, tags, idle timeout, or max duration.
Inputs: session_id (required), name, tags, idle_timeout_minutes, +1 more.
tenki_update_snapshotA
Update a snapshot's mutable metadata (name and/or expiry).
Inputs: snapshot_id (required), name, expires_at.
tenki_update_snapshot_retention_settingsA
Update the workspace's snapshot-retention policy: how long snapshots are kept before automatic cleanup. Pass retention_days to set it, or clear_retention to keep snapshots indefinitely (the unset state) — exactly one of the two. DEPRECATED upstream: prefer tenki_update_workspace_settings, which sets the same retention (and pause retention) alongside the quotas.
Inputs: workspace_id, retention_days, clear_retention.
tenki_update_ssh_keysA
Set the SSH authorized public keys on a running sandbox, enabling direct SSH access for the given keys.
Inputs: session_id (required), public_keys (required).
tenki_update_templateA
Update mutable fields on a template. Only the fields you provide are changed; pass clear_tags to remove all tags.
Inputs: template_id (required), name, tags, env_vars, +8 more.
tenki_update_volumeA
Rename a volume (update its human-readable name). To change a volume's size use tenki_resize_volume instead.
Inputs: name (required), volume_id (required).
tenki_update_workspace_settingsA
Update a workspace's sandbox quotas and retention periods. Only the fields you pass change. Each quota has a matching clear_* flag to remove the override and fall back to the platform default; pass the value OR its clear flag, not both. There are no per-session defaults here — idle timeout and max duration are set per sandbox at creation.
Inputs: max_volumes, workspace_id, max_snapshots, max_templates, +8 more.
tenki_whoamiARead-onlyIdempotent
Return the identity and workspaces for the current API key. Cheap credential test.
Inputs: no named parameters.
tenki_write_fileA
Write a UTF-8 text file to a sandbox (paths under /home/tenki).
Inputs: path (required), content (required), session_id (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 71 tools
Multiple tools overlap significantly: tenki_list_sandboxes and tenki_list_workspace_sandboxes have nearly identical purposes, several snapshot listing variants coexist, and deprecated retention-settings tools duplicate workspace-settings tools. Descriptions often clarify when to use what, but the boundary between run_code and create_sandbox+exec, git and exec, and open_preview versus expose_port/create_preview_url still creates real selection risk.
Every tool uses the tenki_ prefix with a clear verb_noun pattern: create/get/list/update/delete for resources, plus exec/read/write/move for file operations and expose/unexpose for ports. Singular get_ versus plural list_ conventions are consistent, making the API predictable despite its size.
71 tools is far beyond a well-scoped MCP surface and hits the extreme end of the calibration scale. Even accounting for a broad sandbox-management domain, the count is bloated by near-duplicate list/retrieval variants and deprecated endpoints that could be consolidated.
The domain is very thoroughly covered: full sandbox lifecycle, file operations, execution, preview URLs, snapshots, volumes, templates, SSH access, workspace settings, usage billing, and auth diagnostics. There are no obvious dead ends or missing core operations for managing sandboxes and their related resources.
Maintenance
Related MCP Connectors
Cloud Linux VMs for your agent: run code, Docker, files and HTTPS previews in a Sandboxer microVM.
Linux VMs for AI agents: create VMs, run commands, edit files and open preview URLs.
Persistent cloud workspaces for AI agents: run commands, edit files, use git and a browser.
Run AI agent tasks in isolated sandboxes with a cost cap, saved workflows and scoped connections.
Related MCP Servers
- AlicenseAqualityBmaintenanceProvides agents with real Linux VMs to run commands, manage files, expose ports, and track costs, all through an ephemeral sandbox service.11MIT

MileHost MCP Serverofficial
AlicenseCqualityCmaintenanceEnables AI coding agents to manage cloud containers, create and edit files, run commands, and deploy projects directly.247MIT- FlicenseNot gradedqualityDmaintenanceEnables AI agents to operate a persistent Linux sandbox in the cloud, running commands, managing files, using Git, and publishing artifacts through a Streamable HTTP MCP endpoint.-
- AlicenseAqualityBmaintenanceProvides secure, disposable, resource-bounded container sandboxes that let AI agents create and inspect files, run shell commands with timeouts, and perform Git operations without exposing the host environment.11MIT