Code & AI agent sandboxes: read the guide

Celesto AI

Open coreSandbox

Give your agents a Real Computer

Homepage Docs Pricing Repository

Scanned 12 Oct 2026 · Schema v5 · features v21

From: —
Prepaid usage-metered running time, no monthly platform fee; top-ups above $2 (at least $2.01)
Free tier: No
No card required.
Isolation: Virtual machine
hypervisor, or nothing at all
Regions: —
No user-selectable hosted region list or runtime data-residency commitment found. Delaware corporate registration and London headquarters are company facts, not cloud regions. Subprocessor locations are not user workload placement. Self-host placement is operator-controlled.
Capabilities: 25 / 61
8 partial · 23 fail · 0 not assessed
Agent readiness: 11/11 · 10/11
read · drive, pre-login surface

Advertised for

Code execution · Computer use agents · Browser use agents · Coding agents · Long running agents · Untrusted code · Preview apps · GitHub actions CI

Built for

Developers, then technical

· CLI/cloud create+exec+SSH hero; pip/TypeScript SDK + REST/OAuth MCP docs.
· Hardware-isolated microVM/KVM/vsock vocabulary and self-host setup.
· Pricing per fixed CPU/RAM/root-disk running hour.

01 Pricing Prepaid usage-metered running time, no monthly platform fee; top-ups above $2 (at least $2.01)

Tier	EntryTrial	Micro	Small	Standard	Memory	Large	TopXL
Price	One-time $5 runtime credit for 30 days	$0.015/running hour	$0.020/running hour	$0.030/running hour	$0.040/running hour	$0.080/running hour	$0.160/running hour
Includes	—	1 shared vCPU; 2 GB RAM; 20 GB root disk	1 shared vCPU; 4 GB RAM; 25 GB root disk	2 shared vCPU; 4 GB RAM; 30 GB root disk	2 shared vCPU; 8 GB RAM; 40 GB root disk	4 shared vCPU; 16 GB RAM; 60 GB root disk	8 shared vCPU; 32 GB RAM; 100 GB root disk
Storage	—	20 GB	25 GB	30 GB	40 GB	60 GB	100 GB
Compute hours overage	—	$0.015/h	$0.02/h	$0.03/h	$0.04/h	$0.08/h	$0.16/h

Sandbox running hour at selected fixed compute shape·Card required: No·celesto.ai

Overages: No platform minimum; running time consumes prepaid credit, low balance stops sandboxes. Paid up to 10 running; stopped root disks retained free 30 days. No separate published workspace-storage or egress tariff found.

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

Isolation & lifecycle

Isolation primitive

Virtual machinehypervisor, or nothing at all

Persistence

Durable root disk and CelestoFS workspace across stop/start; delete removes state; stopped root disks retained free for 30 days

Snapshots / forks

Yes

Resources & limits

Storage

Included fixed root disks 20/25/30/40/60/100 GB by current size plus separately documented durable CelestoFS workspace filesystem

Limits & quotas

Current pricing: Micro 1 shared vCPU/2 GB RAM/20 GB root → XL 8 shared vCPU/32 GB RAM/100 GB root; paid up to 10 running sandboxes subject to host capacity; trial one retained Micro; publish at most 4 ports (1024–65535)

GPU

Access

Deploy interface

clisdkapi

SDK languages

PythonJavaScriptTypeScript

Network & secrets

Egress controls

Hosted open/off policy fixed at creation; no hosted domain/IP allowlists. Local core separately offers IPv4/CIDR restrictions with backend constraints.

Secrets

Yes

Bring your own cloud

Not documented

Placement & use

Jurisdiction

Not documented

Advertised workloads

Code executionComputer use agentsBrowser use agentsCoding agentsLong running agentsUntrusted codePreview appsGitHub actions CI

03 Capabilities 25 pass · 8 partial · 23 fail · 61 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 7/11

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.

Supported

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.

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 supported

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

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.

Supported

Networking & delivery 4/8

Automatic HTTPSTLS endpoints provisioned automatically per workload.

Supported

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

Supported

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

Not supported

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

Supported

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.

Not 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 6/9

Public APIDocumented programmatic API.

Supported

OpenAPI specMachine-readable OpenAPI / Swagger spec published.

Supported

CLIOfficial command-line tool.

Supported

SDKOfficial client library/libraries.

Supported

SSH accessDirect SSH into workloads.

Supported

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

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 supported

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 3/12

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

Partially supported

Agent skillsInstallable / declarable agent skills or tools.

Partially supported

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

Not supported

MCP serverExposes a Model Context Protocol server for agents.

Supported

MCP clientActs as an MCP client / consumes MCP tools.

Partially supported

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).

Partially supported

BYOK for LLMsBring-your-own LLM provider keys.

Supported

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).

Not 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).

Supported

Data 0/1

First-party databasePlatform-provided managed database.

Not supported

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 2/6

Scheduled tasksCron / scheduled job primitive.

Not supported

Durable workflowsDurable execution / long-running workflow primitive.

Partially 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 supported

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.

Supported

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 3/9

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

Partially supported

SSO / SAMLEnterprise single sign-on.

Not supported

Secrets managementFirst-class secrets storage / injection.

Supported

TeamsMulti-user team accounts.

Partially supported

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

Supported

BYOCRuns inside the customer's own cloud account.

Not supported

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 supported

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 supported

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.

Gatekeeper end-user delegated OAuth connectorsoauth-credential-proxy
“Connect your end users to third-party providers like Google Drive with delegated OAuth access scoped to your project. This is neither general workload login nor a phantom-token egress injection proxy.” celesto.ai
Managed agents with end-user budget caps and exact spend accountingend-user-agent-budgets
“Per-user budgets, exact cost accounting and durable audit trail; records every run against a user and stops spending past the allowed cap.” celesto.ai

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

Can an agent read it?

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

11/11

7 full · 4 partial · 100% of applicable checks

llms.txt
Supported
“/llms.txt 200 substantive site index and /docs/llms.txt 200 rich page index (33811 bytes).”
llms-full.txt
Supported
“/docs/llms-full.txt 200 comprehensive 1004137-byte corpus; date-prefixed POINTER retained. Apex /llms-full.txt 404 does not negate docs copy.”
robots.txt AI policy
Partially supported
“robots.txt names search/social crawlers plus User-agent:* Allow:/; silent about named AI UAs, default-allow.”
JSON-LD
Partially supported
“Homepage has Organization/WebSite/SoftwareApplication JSON-LD graph with name/description, but no offers in the SoftwareApplication record.”
Sitemap
Supported
“Root sitemap 12 URLs; docs XML sitemap 19417 bytes; robots also points to blog sitemap index, successfully rendered after direct 403.”
Meta tags
Supported
“Descriptive title, meta description, OG title/description/url/site/image and Twitter metadata in raw server HTML.”
Markdown negotiation
Supported
“Accept:text/markdown on /docs/cloud/overview returns 200 text/markdown (6156 bytes); homepage remains HTML. Per-page docs .md variants also work.”
Content efficiency
Partially supported
“HTMLParser excludes script/style: 7928 readable-text UTF-8 bytes / 188029 HTML bytes = 4.216%, rubric 1–5% half credit. Homepage is no longer a tiny empty JS shell.”
Docs discoverable
Supported
“Header Docs and hero View docs link to https://celesto.ai/docs; old docs.celesto.ai redirects to the current docs path.”
Machine-readable pricing
Supported
“Dedicated /pricing raw HTML contains parseable six-row CPU/RAM/root-disk/hourly-price table, prepaid/trial/retention copy.”
No anti-bot wall
Partially supported
“Main pages/docs/specs directly fetch; blog and blog sitemap direct requests return Cloudflare error 1010/403, both rescued by sequential foxcrawl 200 (blocked:false). Intermittent/some-path public challenge, not persistent missing content.”

Can an agent drive it?

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

10/11

7 full · 3 partial · 91% of applicable checks

Public API
Supported
“Documented Computers API and managed agents namespaces; cloud SDK/API surface public.”
OpenAPI spec
Supported
“Public GitHub cloud OpenAPI 3.1.0 snapshot with 186 paths and local bridge 19-path OpenAPI spec. API host /openapi.json 404; source repo published specs credited, not a false absent.”
MCP server
Supported
“agents.md documents https://mcp.celesto.ai/mcp with OAuth DCR/PKCE. 401 WWW-Authenticate resource_metadata resolves to public protected-resource metadata and API-host OAuth issuer discovery 200.”
SDK packages
Supported
“Python celesto and TypeScript/JavaScript @celestoai/sdk documented official clients.”
Auth documented
Supported
“Credential lookup order, organization scoping, API-key login/store/logout and OAuth MCP issuer/auth/token/DCR/scope discovery documented.”
Self-serve credentials
Partially supported
“Settings > Security key creation documented; MCP OAuth still requires account browser login. DCR metadata is not proof an autonomous agent can register without a human.”
Webhooks
Not supported
“No documented outbound platform webhook subscription with signing. Exposing webhook receivers and internal Stripe/email/Telegram webhook endpoints do not qualify.”
CLI tool
Supported
“Official celesto CLI local/cloud creates, executes, SSHs and manages computers/ports.”
Error model
Supported
“Documented SDK HTTP error classes/status/data/requestId and managed-agent machine-readable refusal codes plus OpenAPI JSON schemas. Real unauthenticated API/MCP response includes detail/status_code.”
Rate limits documented
Partially supported
“CelestoRateLimitError described for too many requests; managed-session conflict documents Retry-After hint. No numeric general API quota plus rate-limit response headers established.”
Agent config files
Partially supported
“Public /agents.md contains MCP tools/client configuration. Repository AGENTS.md is development guidance, not an independent hosted integration discovery surface; /.well-known/agent returns 404.”

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

Advertised integrations 5

Claude Code Anthropic
Supported
“celesto claude start”
Source github.com
Advertised local-core preset. Pi also has an explicit hosted-cloud extension guide; not a live compatibility test.
Codex CLI OpenAI
Supported
“celesto codex start”
Source github.com
Advertised local-core preset. Pi also has an explicit hosted-cloud extension guide; not a live compatibility test.
OpenCode opencode
Supported
“celesto opencode start”
Source github.com
Advertised local-core preset. Pi also has an explicit hosted-cloud extension guide; not a live compatibility test.
Pi earendil-works
Supported
“celesto pi start”
Source github.com
Advertised local-core preset. Pi also has an explicit hosted-cloud extension guide; not a live compatibility test.
Hermes Agent Nous Research
Supported
“celesto hermes start”
Source github.com
Advertised local-core preset. Pi also has an explicit hosted-cloud extension guide; not a live compatibility test.

Plausible compatibility 10

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.

Aider
Amp
Antigravity
Devin
Droid
Gemini CLI
Goose
Kilo Code
Kiro CLI
Qwen Code

06 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 —

Regions are not disclosed on the public surface.

No user-selectable hosted region list or runtime data-residency commitment found. Delaware corporate registration and London headquarters are company facts, not cloud regions. Subprocessor locations are not user workload placement. Self-host placement is operator-controlled.

Legal

Privacy notice Terms of service

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

Licence & repository

Licence: Apache-2.0OSI
Primary language: Python
Stars: 1,028
Forks: 84
Created: 15 Feb 2026
Archived: No

Activity

Last commit: 11 Oct 2026
Latest release: v0.1.4 · 6 Oct 2026
Issues open / closed: 9 / 57 · 86% closed

v0.1.3 and v0.1.3.post1 on September 28, v0.1.4 on October 6; image releases interleaved. Active pre-1.0 runtime; no maturity grade inferred from stars.

Fetched via gh on 12 Oct 2026.

Self-hosting

Methods: package-managersource-build
You operate: —

Install pip package celesto or build the repository; installer prepares the runtime. README requires Python 3.11+ while install docs say 3.10+; follow current package/README requirements. Linux requires KVM, permissions and host networking tools; macOS uses QEMU/HVF and the macOS guest preview requires Apple Silicon. Firecracker and QEMU are stable backend choices, libkrun experimental. Full snapshots are Linux-guest/isolated-disk scoped and exclude Windows, workspace mounts and extra drives. Network controls have host/backend/port/mount constraints. Local trust-on-first-connection and writable mounts are explicit trust decisions. No external service is required for the basic VM runtime (embedded SQLite); SmolFS/cloud storage is a separate component. Self-host = Apache-2.0 plus own infrastructure; the managed dashboard, hosted billing/CI and control plane are not claimed self-hostable.

Documentation signals 8/8

SupportedQuickstart
SupportedAPI reference
SupportedSelf-host guide
SupportedChangelog
SupportedOpenAPI
Supportedllms.txt
SupportedExamples
SupportedArchitecture doc

08 Company & presence Facts only, not scored

Company

Headquarters: London, United Kingdom
Founded: —
Team size: Not disclosed
Hiring: Not disclosed

Activity

Last release: v0.1.4 (2026-10-06) github.com
Last blog post: Latest visible updated posts: 2026-10-11 (NVIDIA OpenShell comparison and runtime guides) celesto.ai

Presence 8

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

09 Sources & method How this entry was produced

Public surface only

Everything here was read from celesto.ai's unauthenticated pages, docs, pricing and repository on 12 Oct 2026. Nothing behind a login was used.

Statuses, not scores

Capabilities are pass, partial or fail against a written rubric (features v21). “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: v21
Keys tracked today: 61
Assessed on this entry: 61

Something wrong on this page?

Nearby in the atlas

All Sandbox products →