Catalog Systems AI sandbox & runtime environments SandBase Harness

SandBase Harness

https://github.com/sandbaseai/sandbase-harness

A local-first, self-hosted runtime for persistent AI-agent sessions, sandboxed execution, MCP tools, approvals, audit, and replay. **Claims / pitch:** The public repository documents local, Docker, and Kubernetes sandbox backends; persistent sessions and event streams; MCP tools; skills and credential management; approval gates; audit and replay workflows. These are suggestions for curator verification, not endorsement claims. Traction: Submitted via intake issue #173

Active · Verified 2026-10-27 · fresh as of 2026-10-27 · 20 cited facts, 0 with a third-party source · 2 recorded changes · Suggest a correction

At a glance

Type
Harness repo fix
Tier
T3
Section
AI sandbox & runtime environments
Created
2026-07-11 repo fix
Latest release
v0.3.8 (2026-08-30) repo fix
License
Apache-2.0 repo fix
GitHub
sandbaseai/sandbase-harness — ~640 stars, last commit 2026-10 repo fix
Pricing
searched not found fix
Funding
searched not found fix
Freshness
2026-10-27 fix
Layer
agent-infra repo fix
Licensing
open-source repo fix
Deployment locus
local repo fix
Platform vs product
platform repo fix
SDK languages
typescript repo fix
Protocols
mcp-client repo fix

When to use

Optimised for: searched not found

Anti-fit: searched not found

Pros & cons

Pros

searched not found

Cons

searched not found

Claims & capabilities

Local-first, self-hosted AI agent runtime and MCP bridge with sandboxed sessions, memory, credentials, audit/replay, and a local Console.

Technical surface

API surface
searched not found
Backend storage
searched not found
Deployment
Self-hosted only — local-first runtime (local, Docker, Kubernetes or self-hosted worker sandboxes), no hosted control plane
Embedding model
searched not found
Multi-tenancy
searched not found
MCP
yes — MCP toolsets with permission policies/approvals; listed in the official MCP Registry as an MCP bridge
A2A
searched not found
OpenTelemetry
searched not found

Alternatives to SandBase Harness → live options ranked by shared design, licence, hosting and MCP support

Similar systems

Other ai sandbox & runtime environments in the catalog, ranked by inbound references.

  • Daytona T2

    Daytona is secure, elastic sandbox infrastructure for running AI-generated code and agent workflows (isolated sandboxes with snapshots, volumes, computer use, MCP server). Pivoted from an open-source cloud development environment; as of June 2026 core development moved to a private codebase and the daytonaio/daytona repo was archived (2026-10-03). SDKs, API clients and CLI/MCP tools are published in the open daytona/clients repo. Founded by the ex-Codeanywhere team (Ivan Burazin).

  • e2b T1

    e2b provides secure, ephemeral sandboxes (Firecracker microVMs) for AI agent code execution. Each sandbox spins up in <200ms with full Linux environment, Python/Node/etc preinstalled. Substrate for code-agent products including Anthropic Claude Code (artifacts run on e2b), Perplexity, You.com Search, Cognition Devin (partially), and Hugging Face Spaces. Apache 2.0 OSS + managed cloud. $13M Seed + Series A 2024 (Sunflower Capital + a16z).

  • Modal Sandboxes T1

    Modal Labs ships a serverless container platform with first-class GPU support; Sandboxes feature (2024) adds ephemeral, AI-agent-friendly compute sandboxes. Founded 2021 by Erik Bernhardsson (ex-Spotify ML, author of Luigi + Annoy). $87M total raised through Series B (Lux Capital + Redpoint + Definition + Amplify). Used by major LLM apps for batch inference + agent code execution.

  • Anchor (Anchor Browser) T3

    Anchor Browser provides cloud-hosted Chromium instances optimised for AI agent use — anti-detect, persistent sessions, residential proxies, CAPTCHA handling. Founded 2023, YC W24. $2.6M seed disclosed. Distinct from Browserless by adding persistent-state + anti-detect optimised for AI-agent web workflows.

  • Browserless.io T2

    Browserless is a headless-browser-as-a-service — managed Chromium / Playwright / Puppeteer endpoints for web scraping, screenshots, PDF generation, web automation. Increasingly used as substrate for browser-agent products. Founded 2017 by Joel Griffith. Bootstrapped + small seed. Distinct from Playwright OSS — Browserless ships managed Chromium farms with anti-bot circumvention.

  • Coder.com T2

    Coder.com (formerly Coder Technologies) provides enterprise self-hosted cloud development environments — VS Code Server + dev workspaces on K8s / VMs. $30M Series B 2021 (Redpoint + Founders Fund + Coatue). Founded 2017 by Kyle Carberry + Ammar Bandukwala. 2024 launched 'Coder AI' for agent-execution inside workspaces. Self-hostable AGPL OSS + commercial enterprise.

Sources (19 cited facts, 0 third-party)

Each fact links to the page it came from; the badge says who published it (how to read sources).

What changed

Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)

  1. 2026-10-27
    • Latest release “last commit 2026-10-12” → “v0.3.8 (2026-08-30)” source
    • Deployment first recorded as “Self-hosted only — local-first runtime (local, Docker, Kubernetes or self-hosted worker sandboxes), no hosted control plane” source
    • MCP support first recorded as “yes — MCP toolsets with permission policies/approvals; listed in the official MCP Registry as an MCP bridge” source
    #270
  2. 2026-10-23 Added to the catalog #222