Code & AI agent sandboxes: read the guide

Tencent Cloud Agent Runtime (Agent Sandbox / AGS)

HostedSandbox

A new-generation infrastructure platform built around the agent-native execution paradigm — runtime, sandbox, tools, gateway, memory and observability for agents ('用系统的确定性来收敛 Agent 的不确定性').

Homepage Docs Pricing Repository

Scanned 25 Sep 2026 · Schema v5 · features v19

From: —
Pay-as-you-go (postpaid), billed per second, settled hourly: CPU + memory while running, system disk above 15 GiB per instance; mounted storage billed by its own products. Commercial billing since 2026-07-14.
Free tier: No
Card requirement unconfirmed.
Isolation: Kata / gVisor
hardened or user-space kernel
Regions: 10
Beijing, China · Chongqing, China · Guangzhou, China · Hong Kong, China · Nanjing, China · Shanghai, China +4
Capabilities: 20 / 57
10 partial · 22 fail · 4 not assessed
Agent readiness: 9/11 · 9/11
read · drive, pre-login surface

Advertised for

Code interpreter sandboxes for AI generated code · Browser automation / deep research · Computer use and mobile use (Ubuntu desktop, Windows, Android/ADB) agents · Agent RL training and evaluation (SWE bench, OSWorld, Android World, WAA presets) · Coding agents in sandboxes (Claude Code / CodeBuddy Code) · Resident long running agents such as OpenClaw · Custom image services behind a Deployment endpoint

Built for

Developers, then technical

· API/SDK/CLI-first; instances can only be created via API, SDK or MCP
· vocabulary: VM-level isolation, VPC, CAM, QPS, image pre-warm
· usage pricing per core-second

01 Pricing Pay-as-you-go (postpaid), billed per second, settled hourly: CPU + memory while running, system disk above 15 GiB per instance; mounted storage billed by its own products. Commercial billing since 2026-07-14.

Tier	EntryPay-as-you-go (按量计费)
Price	CPU CNY 0.000081/core/s + memory CNY 0.000025/GiB/s + system disk CNY 0.0021/GiB/h above 15 GiB
Includes	All sandbox types and regions at one public price ('不区分地域及可用区'). Launch acceleration (millisecond cold start) is in internal beta, to be billed later; pause-period storage not yet charged while pause is in beta.
Storage	15 GB
Storage overage	CN¥1.512/GB
vCPU hours overage	CN¥0.2916/h
Memory gb hours overage	CN¥0.09/h

core-second (CPU) + GiB-second (memory) + GiB-hour (system disk above 15 GiB)·Card required: unconfirmed·cloud.tencent.com

Overages: Pure metering; nothing included except 15 GiB disk per instance. External storage (COS, CFS, AgentBucket) billed by those products.

02 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

Ephemeral by default: each instance has a lifecycle (default 5 min, max 24 h per the tool docs; SDK default 300 s) and its system disk is deleted with it. Pause/resume keeps state; Deployments keep instances behind a stable endpoint with idle PAUSE/STOP. Durable data lives on mounted COS/CFS/AgentBucket.

Snapshots / forks

Pause/resume with process-level snapshots; no live fork

Resources & limits

Storage

Per-instance system disk (1–20 GiB depending on type; 15 GiB free per instance, then CNY 0.0021/GiB/h) + mounts: COS, CFS, image volumes, AgentBucket; TurboCFS and JuiceFS in grey release. Up to 10 mounts per tool.

Limits & quotas

Per account per region (raisable): 10 sandbox tools, 50 instances (running + idle), 50 vCPU and 100 GiB running, 20 paused. 3 API keys by default. Instance lifecycle max 24 h (tool docs). API QPS 20–100 per action.

GPU

Access

Deploy interface

apisdkclimcpconsole

SDK languages

PythonGoJavaScript/TypeScript (via E2B SDK compatibility)

Network & secrets

Egress controls

Yes

Secrets

Identity Credentials module (OAuth2 / API keys for external tools), CAM roles (RoleArn) for cloud resources, env vars in tool config

Bring your own cloud

Placement & use

Jurisdiction

China (Tencent Cloud China-site service, PRC entities); regions also in Hong Kong, Singapore, Tokyo, Frankfurt and Ashburn

Advertised workloads

Code interpreter sandboxes for AI generated codeBrowser automation / deep researchComputer use and mobile use (Ubuntu desktop, Windows, Android/ADB) agentsAgent RL training and evaluation (SWE bench, OSWorld, Android World, WAA presets)Coding agents in sandboxes (Claude Code / CodeBuddy Code)Resident long running agents such as OpenClawCustom image services behind a Deployment endpoint

03 Capabilities 20 pass · 10 partial · 22 fail · 57 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 5/10

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

Partially 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.

Partially 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).

Partially 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.

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.

Not assessed

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.

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 5/9

Public APIDocumented programmatic API.

Supported

OpenAPI specMachine-readable OpenAPI / Swagger spec published.

Not supported

CLIOfficial command-line tool.

Supported

SDKOfficial client library/libraries.

Supported

SSH accessDirect SSH into workloads.

Partially 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.

Partially 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).

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 3/11

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

Not supported

Agent skillsInstallable / declarable agent skills or tools.

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).

Supported

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

Not supported

BYOK for LLMsBring-your-own LLM provider keys.

Not 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).

Partially 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 0/1

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 1/6

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 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.

Not 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 2/7

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

Partially supported

SSO / SAMLEnterprise single sign-on.

Not supported

Secrets managementFirst-class secrets storage / injection.

Partially supported

TeamsMulti-user team accounts.

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 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.

Not 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.

Mobile / Android sandboxes with ADBmobile-device-sandbox
“mobile … 预置 Mobile 运行环境和 ADB 能力; android-world … Android World 镜像环境” cloud.tencent.com
Agent / Skills / MCP registry with approval workflowagent-registry
“ApproveRegistryRecord 审批通过 Record 版本 … GetSkillPackageUploadURL 生成 Skill 包上传 URL” cloud.tencent.com
Preset agent-benchmark environments (SWE-bench, OSWorld, Android World, WAA)benchmark-environments
“swebench … 软件工程智能体评测; osworld … 桌面级交互; waa … Windows Agent Arena” cloud.tencent.com

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

Can an agent read it?

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

9/11

5 full · 4 partial · 82% of applicable checks

llms.txt
Partially supported
“cloud.tencent.com/llms.txt exists (8.4 KB, site-wide product index) but does not link Agent Runtime/AGS or its docs.”
llms-full.txt
Partially supported
“cloud.tencent.com/llms-full.txt exists (3.3 MB) but is developer-community articles; no AGS content (artifacts/cloud.tencent.com_llms-full.txt.POINTER.md).”
robots.txt AI policy
Partially supported
“robots.txt is silent on GPTBot/ClaudeBot/etc. (only YisouSpider is fully disallowed) — default-allow.”
JSON-LD
Not supported
“No application/ld+json on the AGS product page.”
Sitemap
Supported
“robots.txt → https://cloud.tencent.com/sitemap/sitemap.xml, a sitemap index of 532 child sitemaps.”
Meta tags
Partially supported
“Title 'Agent Runtime_Agent沙箱_Agent网关- 腾讯云' + description + keywords; no Open Graph tags.”
Markdown negotiation
Not supported
“Accept: text/markdown returns HTML; <url>.md returns the HTML shell.”
Content efficiency
Supported
“Product page is server-rendered: readable text ≈13% of HTML bytes.”
Docs discoverable
Supported
“'产品文档' links to /document/product/1814 in the static hero and help section.”
Machine-readable pricing
Supported
“Billing doc (/document/product/1814/133249) has structured unit-price tables (CPU/core/s, memory/GiB/s, disk/GiB/h) and worked examples.”
No anti-bot wall
Supported
“cloud.tencent.com product and docs pages fetch without a challenge. The international site (tencentcloud.com) served a JS challenge to curl, but it has no AGS page anyway.”

Can an agent drive it?

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

9/11

6 full · 3 partial · 82% of applicable checks

Public API
Supported
“Tencent Cloud API 3.0 `ags` service (API overview 1814/124833) + E2B-compatible data plane.”
OpenAPI spec
Not supported
“No OpenAPI/Swagger file; APIs are documented as Tencent Cloud API 3.0 pages.”
MCP server
Supported
“Documented open-source MCP servers for AGS (github.com/TencentCloudAgentRuntime/ags-mcp: browser-mcp + code-mcp, with setup docs); docs list MCP as an access method.”
SDK packages
Supported
“Python and Go SDKs; E2B SDK compatibility; Tencent Cloud API SDKs in many languages for the control plane.”
Auth documented
Supported
“API key docs (1814/123826), TC3 signature/common parameters (api 1814/124827), sandbox access tokens and signed URLs (1814/132313).”
Self-serve credentials
Partially supported
“CreateAPIKey exists as an API, but it needs Tencent Cloud account credentials; the quickstart creates keys in the console, behind a human Tencent Cloud signup.”
Webhooks
Not supported
“No event webhooks documented (NetworkPolicy 'notification hooks' are traffic-policy callbacks in grey release, not lifecycle webhooks).”
CLI tool
Supported
“Official `agr` CLI (github.com/TencentCloudAgentRuntime/ags-cli).”
Error model
Supported
“Error-code page (api 1814/124832) and FAQ error codes; E2B-compatible errors carry RequestId (changelog 2026-06-18).”
Rate limits documented
Partially supported
“Per-action QPS limits and RequestLimitExceeded documented (1814/123815); no rate-limit response headers documented.”
Agent config files
Partially supported
“One surface: an installable `ags` Agent Skill in the cookbook (plus `agr schema` machine-readable command schema). No agents.md or /.well-known agent descriptor.”

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

Advertised integrations 2

Claude Code Anthropic
Supported
“Coding Agent：AI 生成的代码在云端沙箱中运行 … 支持 CodeBuddy Code / Claude Code 等沙箱环境。”
Source cloud.tencent.com
Medium confidence: a use-case claim, not a dedicated integration guide.
OpenClaw Peter Steinberger
Supported
“openclaw-cookbook | Node.js + custom image + COS | Run OpenClaw in AGS with official image”
Source github.com
Also the product page: resident mode carries 'OpenClaw' long-running jobs ('常驻式运行，承载“OpenClaw”龙虾类长时作业').

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.

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 10

Beijing, Chinaap-beijing
Chongqing, Chinaap-chongqing
Guangzhou, Chinaap-guangzhou
Hong Kong, Chinaap-hongkong
Nanjing, Chinaap-nanjing
Shanghai, Chinaap-shanghai
Singaporeap-singapore
Tokyo, Japanap-tokyo
Frankfurt, Germanyeu-frankfurt
Ashburn (Virginia), USna-ashburn

The API Region parameter's allowed values (10 regions). Tool types vary by region (e.g. Android World only in Beijing and Shanghai per the changelog). Data-plane domain per region: {region}.tencentags.com. Financial-zone endpoints (Shanghai/Shenzhen FSI) exist for the API but are not in the Region list.

Legal

Privacy notice Terms of service

07 Company & presence Facts only, not scored

Company

Headquarters: Shenzhen, China
Founded: Tencent Cloud brand: public-cloud timeline on its about page starts in 2013; Agent Sandbox docs date from 2025-09 (first changelog month)
Team size: Not disclosed
Hiring: not checked

Activity

Last release: Changelog latest entry 2026-06-25 (page updated 2026-07-10); docs pages updated to 2026-09-15; agr CLI repo pushed 2026-09-22 (v0.6.x) cloud.tencent.com
Last blog post: n/a — no AGS blog

Presence 3

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 cloud.tencent.com'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 v19). “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: v19
Keys tracked today: 61
Assessed on this entry: 57

Something wrong on this page?

Nearby in the atlas

All Sandbox products →