Code & AI agent sandboxes: read the guide

Agent Substrate

Open source toolSandbox

A secure-by-default agent execution runtime engineered to run millions of sandboxes with 10x higher density than standard container runtimes.

Homepage Docs Repository

Scanned 25 Sep 2026 · Schema v5 · features v18

Licence: Apache-2.0
Self-hosted; you pay for the infrastructure.
Free tier: Free to run
Isolation: Kata / gVisor
hardened or user-space kernel
Regions: —
Not disclosed on the public surface.
Capabilities: 10 / 55
8 partial · 22 fail · 6 not assessed
Agent readiness: 4.5/7 · 5/9
read · drive, pre-login surface

Advertised for

Stateful AI agents (ADK, LangChain) · Coding agents (Claude Code, Codex, Antigravity) · Ephemeral sandboxes for untrusted LLM generated code · Sandboxed MCP servers · RL rollouts and agent benchmarks · Stateful services

Built for

Developers, then technical

· platform-engineer vocabulary: Kubernetes, WorkerPools, gVisor, SPIFFE, CSI, nftables
· onboarding is `hack/create-kind-cluster.sh` + `install-ate-kind.sh` + kubectl plugin
· deep reference docs and a threat model, no GUI

01 Runtime Each fact carries its source; dots show confidence. Open a row for the evidence.

Isolation & lifecycle

Isolation primitive

Kata / gVisorhardened or user-space kernel

Persistence

Actor RAM + root filesystem are snapshotted to object storage on suspend and restored on resume (possibly on another worker); CSI external volumes persist separately and are not part of snapshots. An evicted actor gets 30 minutes to suspend, otherwise work since the last snapshot is lost.

Snapshots / forks

Suspend/resume full-state snapshots (sub-500ms resume claimed), RevertActor to the last snapshot, tags (atespace-scoped or PUBLISHED) and new actors seeded from a tag; golden snapshots per ActorTemplate.

Resources & limits

Storage

Snapshots in GCS (S3-compatible via rustfs in the kind install); per-actor CSI external volumes provisioned from the ActorTemplate (no PV/PVC objects). Filestore agent volumes (NFS, RWX) optional on GKE.

Limits & quotas

Per-actor CPU/memory right-sizing in the ActorTemplate; worker pools sized by the operator (HPA supported). Design claims: millions of sandboxes per cluster, sub-500ms resume, 500+ suspend/resume per second, 30x oversubscription. 30-minute eviction grace period. IPv4-only actor networking.

GPU

none (device passthrough temporarily removed)

Access

Deploy interface

cliapi

SDK languages

Network & secrets

Egress controls

All actor TCP egress except DNS is tunnelled (mTLS CONNECT) to an egress policy enforcement point; only HTTP(S) 1.1/2 is allowed, WebSocket/CONNECT/other TCP and non-DNS UDP are blocked. Since v0.2.0 a per-actor EgressPolicy (hostname and CIDR rules) is enforced default-deny. Not supported on GKE's build yet.

Secrets

Credential injection at the egress gateway (MITM with a projected trust bundle) resolving ate-secret:// references from Kubernetes Secrets, with a default-deny atespace→namespace policy; per-actor SPIFFE certs and JWTs via the ActorIdentity service.

Bring your own cloud

Placement & use

Jurisdiction

Advertised workloads

Stateful AI agents (ADK, LangChain)Coding agents (Claude Code, Codex, Antigravity)Ephemeral sandboxes for untrusted LLM generated codeSandboxed MCP serversRL rollouts and agent benchmarksStateful services

02 Capabilities 10 pass · 8 partial · 22 fail · 55 of 61 assessed

Supported — Documented on the public surface and verified.
Partially supported — Exists with caveats: gated, thin, or enterprise-only.
Not supported — Checked and not offered.
Not assessed — Scanned before this capability was tracked. Not a lack.

Compute & storage 3/8

Persistent diskStorage that survives restart/stop (not ephemeral-only).

Supported

Object storageManaged S3-compatible / blob object store the platform PROVIDES to a user's workloads (buckets, presigned URLs). DISTINCT from persistent-disk (block/volume attached to an instance) and from a self-host dependency (source.self_host.deps[]=["object-storage"] is an infra dep to RUN an OSS tool — the opposite direction). SCOPE varies and matters: native first-class (Railway Buckets, Cloudflare R2) vs a sibling SKU / extension (Fly Tigris, Bunny Storage, DigitalOcean Spaces) — record which in the evidence quote / note. Admitted v15.

Not supported

Snapshots / forkingSnapshot or copy-on-write fork of a running instance (filesystem-only counts). Whether the snapshot also restores memory + running processes is recorded separately by memory-snapshot (v19).

Supported

Live-state (memory) snapshotsA snapshot / checkpoint / pause captures a running instance's MEMORY and PROCESS state and restores it, so running processes continue where they left off instead of a freshly booted instance with the same files. Counts whether user-triggered (snapshot, checkpoint, pause-to-storage) or used to restore pre-initialised state (memory snapshots for fast cold starts). Distinct from snapshots-forks (any snapshot or fork, including filesystem-only; a product passing that key may still fail this one) and from scale-to-zero (an idle-cost property; suspension may or may not keep memory). Actor multiplexing (see scale-to-zero) records its live-state half here — only when the suspended actor's MEMORY/process state is what comes back, not a freshly booted worker with the actor's files.

Not assessed

Scale to zeroIdle instances suspend at no / negligible cost. ACTOR MULTIPLEXING (many suspended stateful actors sharing a small pool of pre-warmed workers, woken on request — Agent Substrate, AX, kagent; Rivet's hibernating actors) has NO key of its own (proposal REJECTED v20) — score the idle-cost half HERE and the live-state half under memory-snapshot; the oversubscription ratio / density is a performance characteristic the registry does not key.

Partially supported

GPU computeGPU-backed instances available.

Not supported

Multi-regionCompute placeable in more than one geographic region.

Not supported

Sandbox templates / custom imagesBuild reusable sandbox/VM images from an OCI image or Dockerfile (a base you fork instances from).

Supported

Cloud-init / ignition configCustomise a VM / microVM / instance at boot via cloud-init (or ignition) user-data — the same boot-metadata interface as AWS/Azure/GCP. Inject packages, users, SSH keys, files, and run-commands declaratively at launch. Distinct from sandbox-templates (a PREBUILT image you fork instances from) — this is PER-LAUNCH declarative customisation, not a baked image.

Not applicable

Desktop / computer-use sandboxA virtual desktop (GUI) for LLM computer-use — browser + desktop automation, often via noVNC/VNC.

Not supported

Selectable isolation runtimeThe user chooses the isolation runtime per workload / environment / sandbox from more than one option (e.g. microVM vs gVisor vs plain container, native namespaces vs gVisor, gVisor vs VM). Records the CHOICE; the isolation_primitive facet still records the best-evidenced primary. A platform that switches runtime automatically without a user choice (e.g. gVisor forced for GPU) does not pass on that alone.

Not assessed

Networking & delivery 2/5

Automatic HTTPSTLS endpoints provisioned automatically per workload.

Not applicable

Built-in reverse proxyPlatform routes external traffic to workload services.

Supported

Custom domainsBring-your-own domain via CNAME/DNS.

Not applicable

Public preview URLsShareable per-deploy / per-branch URLs.

Not applicable

Direct public IPv6Per-instance stable public IPv6 for raw TCP/UDP (beyond a managed HTTPS proxy) — e.g. mosh / raw sockets.

Not supported

Egress controlsConfigurable OUTBOUND network policy — egress firewall / allowlist / egress proxy on traffic LEAVING a workload. Scoped to OUTBOUND only (v15): private-network REACHABILITY between a user's own services is private-networking, not this. Static-egress-IP belongs here (an outbound concept).

Supported

Private networkingA platform-internal private network between a user's OWN services — service-to-service connectivity, VPC, service mesh, and/or service discovery, not exposed to the public internet. DISTINCT from egress-controls (OUTBOUND firewall/allowlist), reverse-proxy (PUBLIC inbound ingress), and direct-ipv6 (per-instance PUBLIC addressing). Admitted v15.

Partially supported

Email send / receiveSend and/or receive email from the workload.

Not supported

Edge KV storeA globally-replicated key-value store read from WITHIN an edge runtime at the PoP (low-latency, eventually-consistent) — Cloudflare Workers KV, Deno KV, Fastly KV Store, Akamai EdgeKV, Vercel Edge Config. Distinct from first-party-db (a managed DATABASE service you provision and connect to) and from http-query-api (HTTP access to an external store) — edge-kv is colocated with the edge compute and read at the edge.

Not applicable

Developer access 2/8

Public APIDocumented programmatic API.

Supported

OpenAPI specMachine-readable OpenAPI / Swagger spec published.

Not supported

CLIOfficial command-line tool.

Supported

SDKOfficial client library/libraries.

Partially supported

SSH accessDirect SSH into workloads.

Partially supported

Filesystem APIRemote file operations over the API (list/read/write/watch/copy/upload) without SSH.

Partially supported

Git deployDeploy from a Git repo / push-to-deploy.

Not supported

Private container registryHosted private image registry.

Not supported

E2B-compatible APIDrop-in compatibility with the E2B sandbox API/SDKs — existing E2B client code runs against this product by changing the endpoint/domain + key, with no business-code changes. E2B itself passes as the reference implementation. Distinct from public-api / sdk (the product has its own API — this key says it also speaks another vendor's) and from self-host-oss (a self-hosted E2B-compatible server is scored on both).

Not assessed

HTTP data APIQuery / read / write the stored DATA over plain HTTP/REST (or GraphQL) WITHOUT the native protocol driver (Postgres wire / Redis RESP) — agent- and edge-friendly data access via fetch(). The DATA plane; distinct from public-api (a control-plane / management API). Auto-generated-from-schema REST/GraphQL (PostgREST, pg_graphql) is a flavour of this.

Not applicable

AI & agents 1/9

Built-in coding agentShips a first-party AI coding agent in the product.

Not supported

Agent skillsInstallable / declarable agent skills or tools.

Not supported

AI app builderPrompt-to-app / vibe-coding generation surface.

Not supported

MCP serverExposes a Model Context Protocol server for agents.

Partially supported

MCP clientActs as an MCP client / consumes MCP tools.

Not applicable

MCP-server hostingRun/host MCP servers as a workload on the platform (distinct from exposing a first-party MCP control server — that is mcp-server).

Partially supported

LLM gatewayRoutes LLM API calls through the platform (multi-provider).

Not supported

BYOK for LLMsBring-your-own LLM provider keys.

Not applicable

Model inference API (first-party)A FIRST-PARTY hosted LLM/model inference API — the provider serves its OWN catalogue of (usually open-weight) models behind a (typically OpenAI-compatible) endpoint, billed per token. Distinct from llm-gateway (a multi-PROVIDER proxy/router for YOUR calls to external providers like OpenAI/Anthropic) and from byok-llm (bring your own provider keys) — here the platform IS the inference provider, serving its own models.

Not supported

Autonomous QA agentA first-party AUTONOMOUS AI agent that QA-TESTS a running app / implementation — exercising it like a real user (clicking through flows, checking behaviour) and reporting bugs/defects (often back to a coding agent, or before a human review). Distinct from built-in-agent (a code-AUTHORING agent — this one VERIFIES the result, it does not write the feature; a product can ship both), from ai-guided-code-review (organises a diff for a HUMAN to review — not an autonomous tester), and from observability (runtime metrics/logs, not active testing).

Not supported

Credential-injection proxyThe platform holds real secrets OUTSIDE the workload and injects them on the wire (an egress / broker proxy swaps a placeholder or phantom token for the real credential on matching outbound requests), so code running in the sandbox — e.g. an untrusted agent — never sees or can exfiltrate the real key. Distinct from secrets-management (storage + injection INTO the workload as env vars/files — the workload holds the secret) and from egress-controls (outbound allow/deny POLICY; a product can have an allowlist without injecting credentials, and vice versa).

Supported

Preinstalled coding agentThe platform installs and launches a THIRD-PARTY coding agent (a harnesses.yml entry — Claude Code, Codex, OpenCode, Pi ...) inside a workspace/sandbox it provisions, with the agent's auth wired in (your own account or API key), so you start an agent session rather than an empty box. Distinct from built-in-agent (a FIRST-PARTY agent the vendor ships) and from harnesses[] (which records that a harness CAN target the product, not that the product runs it for you).

Not assessed

Data 0/0

First-party databasePlatform-provided managed database.

Not applicable

Database branchingCopy-on-write branch of a DATABASE (schema + data) as a dev-lifecycle primitive — branch your data like code, usually instant via a metadata operation. Distinct from snapshots-forks (forks a running compute INSTANCE / VM state) and from preview-environments (a full per-PR environment that may USE a DB branch). For a data product a CoW branch maps HERE, not to snapshots-forks.

Not applicable

Vector searchFirst-class storage, indexing, and similarity search of vector embeddings — a native vector type/index or a managed vector DB. The data-layer primitive for RAG / semantic retrieval.

Not applicable

Ephemeral instant DB (no-signup)Provision a working, usable instance via a single command / HTTP call with NO account, signup, or card — typically time-boxed (expires unless claimed). A zero-onboarding path an agent can hit directly.

Not applicable

Platform & operations 1/4

Scheduled tasksCron / scheduled job primitive.

Not supported

Durable workflowsDurable execution / long-running workflow primitive.

Not supported

Preview environmentsA full EPHEMERAL environment spun up per PR / branch — compute + addons + (often) a DB branch — torn down on merge. More than preview-urls (which is just a shareable URL for one deploy).

Not applicable

Built-in observabilityFirst-party metrics / logs / tracing / APM / profiling included in the platform (no third-party stack required to see what a workload is doing).

Supported

Managed CI/CD runnersPlatform-hosted, drop-in CI/CD runner VMs (GitHub Actions / GitLab), typically faster and cheaper than the CI vendor's own hosted runners, billed per runner-minute; isolated + ephemeral per job. Distinct from git-deploy (push-to-deploy of an APP) and scheduled-tasks (cron) — this is managed CI COMPUTE you point your existing pipelines at.

Not applicable

Project template marketplaceA marketplace / gallery of prebuilt PROJECT STARTER templates (apps or websites) a user clones or launches a new project from and then customises — the "start from a template" surface of a builder / PaaS. Distinct from sandbox-templates (a COMPUTE base IMAGE you fork runtime INSTANCES from — Dockerfile / OCI), from app-builder (the prompt-to-app GENERATION surface — a product can have both), and from one-click OSS-app catalogues / helm-addon-marketplace (deploy a FINISHED app as-is, not a starter you build on).

Not supported

Organisation & security 1/6

Built-in authPlatform-provided authentication / IAM for workloads.

Not supported

SSO / SAMLEnterprise single sign-on.

Not supported

Secrets managementFirst-class secrets storage / injection.

Partially supported

TeamsMulti-user team accounts.

Not supported

SharingShare a running workload/resource via link (optionally gated).

Not supported

BYOCRuns inside the customer's own cloud account.

Not applicable

Deploy into customers' VPCsA SaaS vendor uses the platform to deploy and operate ITS software inside ITS CUSTOMERS' cloud accounts / VPCs (nested or "customer BYOC"), managed from the vendor's control plane. Distinct from byoc (your own workloads in your own cloud account).

Not assessed

Workload identityThe platform gives each workload its own verifiable identity (OIDC token, SPIFFE SVID, cloud IAM federation) that it exchanges for short-lived credentials, so no long-lived static secret is stored or injected. Distinct from secrets-management (stores / injects static secrets) and credential-injection-proxy (swaps real secrets in on the wire).

Not assessed

Self-host / OSS coreOpen-source core that can be self-hosted.

Supported

Proposed, not verified

Capabilities the scan saw but the registry does not track yet. A key is admitted only once two distinct products exhibit it. These are not scored.

Actor multiplexing: many suspended stateful actors share a small pool of pre-warmed workers, woken on request (30x+ oversubscription claim)actor-multiplexing
“Demonstrates 30x+ oversubscription by "juggling" a large registry of stateful actors onto a small pool of shared physical pods.” github.com
Request parking: router holds inbound requests during worker-pool saturation instead of returning 503request-parking
“lets the `atenet` router hold ("park") an inbound request whose target actor cannot be served *yet* because of transient worker-pool saturation” github.com
Workload identity: per-actor SPIFFE mTLS certs and OIDC JWTs that survive migration between workersworkload-identity
“Workloads can exchange their ephemeral Kubernetes credentials for stable **Actor Identity** credentials that persist even as the process migrates between different physical workers.” github.com

03 Agent readiness Pre-login surface · counts, not a grade

Can an agent read it?

parse-ability · docs, metadata, machine-readable surfaces

4.5/7

4 full · 1 partial · 64% of applicable checks

llms.txt
Not supported
“No llms.txt: no own site; learn.agentsubstrate.dev/llms.txt → 404”
llms-full.txt
Not supported
“No llms-full.txt anywhere (learn.agentsubstrate.dev/llms-full.txt → 404)”
robots.txt AI policy
Not applicable
“No own marketing apex (ate.dev redirects to GitHub)”
JSON-LD
Not applicable
“No own site”
Sitemap
Not applicable
“No own marketing apex”
Meta tags
Partially supported
“GitHub repo description 'Agent Substrate: the core system' (thin); no topics set; GKE docs page carries full metadata”
Markdown negotiation
Supported
“README + docs/*.md are native markdown, raw-fetchable from raw.githubusercontent.com (HTTP 200)”
Content efficiency
Supported
“Markdown repo docs; no JS shell”
Docs discoverable
Supported
“README 'Documentation & Guides' links architecture, API guide, CLI, glossary, observability, auth, egress, threat model, roadmap”
Machine-readable pricing
Not applicable
“No pricing (oss-tool, self-host)”
No anti-bot wall
Supported
“GitHub + Google docs fetched with no challenge”

Can an agent drive it?

act-ability · APIs, SDKs, MCP, credentials, webhooks

5/9

3 full · 4 partial · 56% of applicable checks

Public API
Supported
“docs/api-guide.md §6: ateapi.Control gRPC (Create/Update/Resume/Suspend/Revert/Delete/Get/List) + §7 ActorIdentity”
OpenAPI spec
Not supported
“gRPC/protobuf only; no OpenAPI”
MCP server
Partially supported
“No MCP server for the core control plane; the first-party agent-substrate/env add-on (alpha) documents a per-environment MCP endpoint (POST /v1alpha/envs/{id}/mcp) with exec/file tools”
SDK packages
Partially supported
“Core: Go types in-repo only; the separate env add-on adds Go + Python clients (not for the core API)”
Auth documented
Supported
“docs/authentication.md: mTLS client certs + bearer JWT providers config; ActorIdentity MintJWT/MintCert caller rules”
Self-serve credentials
Not applicable
“Self-hosted: there is no vendor-issued credential to obtain”
Webhooks
Not supported
“No webhooks; lifecycle events are emitted over OTLP”
CLI tool
Supported
“Official kubectl-ate plugin”
Error model
Partially supported
“gRPC status codes documented per RPC (INVALID_ARGUMENT, ABORTED, FAILED_PRECONDITION, PERMISSION_DENIED), no general error-model doc”
Rate limits documented
Not applicable
“Self-hosted API with no vendor-imposed rate limits”
Agent config files
Partially supported
“AGENTS.md (+ CLAUDE.md) at the repo root and .agents/skills/, but aimed at contributors to Substrate, not at agents operating it; no /.well-known/agent”

04 Coding harnesses Which agents and IDEs are documented to run here

Advertised integrations 3

Claude Code Anthropic
Supported
“A demo of three Claude-Code-driven agents sharing two Agent Substrate pods.”
Source github.com
Codex CLI OpenAI
Supported
“Claude Code, CodeX, and Antigravity: Support for high-density, stateful coding environments that preserve system state and filesystem state across sessions.”
Source github.com
Hermes Agent Nous Research
Supported
“Nous Research, the team behind the Hermes Agent, is actively building on top of Agent Substrate.”
Source cloud.google.com

Plausible compatibility 0

Derived, not stored: the product passes SSH access, so any CLI coding agent is installable inside it. This is plausible compatibility, not a supported integration.

No general shell access documented, so nothing is derived. Absence is not unsupported.

05 Compliance & regions Standards held, with ordinal rigor; never a score

Certifications held 0

No attestations found on the public surface. Empty means nothing is advertised, after checking, not that nothing was checked.

Regions —

No region list captured.

06 Open source Apache-2.0 · fetched by machine, not quoted from marketing

Licence & repository

Licence: Apache-2.0OSI
Primary language: Go
Stars: 3,802
Forks: 443
Created: 13 May 2026
Archived: No

Activity

Last commit: 25 Sep 2026
Latest release: v0.2.0 · 25 Sep 2026
Issues open / closed: 333 / 309 · 48% closed

3 releases: v0.0.0 (2026-05-19, launch), v0.1.0 (2026-09-09), v0.2.0 (created 2026-09-25 17:29, published 22:26). Release notes are detailed with a Breaking Changes section; no binary assets (build/deploy from source with ko).

Fetched via gh on 25 Sep 2026.

Self-hosting

Methods: kubernetessource-build
You operate: kubernetespostgresobject-storage

Self-host is the only way to run it: free under Apache-2.0 + your own Kubernetes cluster (latest stable minor + previous). Needs object storage for snapshots (GCS on GKE; rustfs/S3-compatible in the kind install) and PostgreSQL for control-plane state (Cloud SQL option on GCP). Installed from source with hack/install-ate.sh (ko-built images, a pullable registry); no Helm chart. Required beta APIs (PodCertificateRequest, ClusterTrustBundle) can only be enabled at cluster creation on GKE; GKE Standard 1.36+ only, no mixed-arch E2 machines, Workload Identity required. microVM pools need /dev/kvm (nested virtualization). No GPU passthrough. Pre-1.0: breaking changes each release. GKE's supported build lacks EgressPolicy.

Documentation signals 6/8

SupportedQuickstart
SupportedAPI reference
SupportedSelf-host guide
SupportedChangelog
Not supportedOpenAPI
Not supportedllms.txt
SupportedExamples
SupportedArchitecture doc

07 Company & presence Facts only, not scored

Company

Headquarters: Mountain View, CA, USA (Google)
Founded: 2026
Team size: Not disclosed
Hiring: Unknown

Notable people

Tim Hockin (thockin): Kubernetes co-founder, maintainer and co-author of the GKE launch post
Jaana Dogan (rakyll): maintainer, lead author of AX
Benjamin Elder (BenTheElder): kind maintainer, top committer
Solo.io maintainers (Eitan Yarmush, Keith Mattix, Krisztian Fekete) from the kagent side

Activity

Last release: v0.2.0 — 2026-09-25 github.com
Last blog post: 2026-09-15 — 'Agent Substrate brings high-density, scalable, trusted infrastructure to GKE' cloud.google.com

Presence 7

Links found on the public surface scanned; not an exhaustive search.

08 Sources & method How this entry was produced

Public surface only

Everything here was read from github.com/agent-substrate/substrate's unauthenticated pages, docs, pricing and repository on 25 Sep 2026. Nothing behind a login was used.

Statuses, not scores

Capabilities are pass, partial or fail against a written rubric (features v18). “Not assessed” means the scan predates the key.

Confidence is separate

Each fact carries its own confidence. Low-confidence facts are shown, never hidden, so you can weigh them.

Schema: scan v5
Features registry: v18
Keys tracked today: 61
Assessed on this entry: 55

Something wrong on this page?

Nearby in the atlas

All Sandbox products →