Runtime: Sandboxed Coding Agents for Every Team Member
Runtime (YC P26) is a platform that lets your whole team (including non-engineers) use coding agents like Claude Code, Codex, Cursor, Copilot, Gemini, and Devin without needing engineering to handhold every session. It was built by Gus and Carlos, who experienced the pain of rolling out agents across a team: PRs were unmergeable slop, every repo required local setup, skills lived in one person's head, and there was no safe way for a PM to touch a real codebase without risking a bad deploy or secret leak.
Key Features
- Snapshotted Environments: Engineers define context once (system instructions, skills, scoped integrations) and Runtime snapshots the full running environment (multi-service Docker Compose, Kafka, Redis, seeded DBs). Every session boots in milliseconds with all servers already running.
- Sandbox Orchestration: Runs across E2B, Daytona, EC2, or self-hosted K8s. Secrets are injected via a managed proxy so agents never touch them directly.
- Guardrails at Infrastructure Level: Command allow/deny lists, network egress controls, RBAC scoped per human and per agent. Spend limits, file rules, and approval gates built in.
- Shareable Preview URLs: Every session gets a live preview URL so internal builds go from sandbox to the team without production access.
- Multichannel Triggers: Sandboxes can be triggered from the web app, CLI, Slack, Linear, GitHub, or API.
- Specialized Agents: Build agents that run in background (e.g., an alert inspector that wires PagerDuty, Sentry, and repo to find causes and open PRs with unit tests before anyone is paged; a finance agent in private Slack that pulls from Stripe, NetSuite, and Snowflake for reconciliations).
Pricing
Flat platform fee plus compute — no token markup. Free tier available at app.runtm.com.
Open Source
The core is open source at github.com/runtm-ai/runtm under MIT (templates), Apache 2.0 (CLI & shared libs), and AGPL v3 (API & worker). Self-hostable on your own infrastructure.
Who It's For
Platform teams wanting to roll out coding agents safely across engineering, product, design, marketing, support, finance, and people teams.
📖 Read the full source: HN LLM Tools
👀 See Also
ClawVibe: A Hands-Free iOS Voice Assistant for AI Agents with On-Device STT/TTS
ClawVibe is a native iOS app that provides hands-free voice interaction with AI agents during commutes. It uses on-device speech recognition and TTS, supports CarPlay, and includes voice biometrics to filter background noise. Only text is sent over the network.
OpenClaw Optimizer v1.18.0 released with OpenClaw v2026.3.7 alignment
OpenClaw Optimizer skill v1.18.0 is now aligned with OpenClaw v2026.3.7, adding support for new AI providers including Google Gemini 3.1 Flash-Lite and OpenAI gpt-5.4, plus new CLI commands like /session idle and /usage cost.
Git pre-commit hook prevents AI coding agents from committing with stale documentation
A developer created a Git pre-commit hook that blocks commits when documentation files are outdated, specifically addressing issues with AI coding agents like Claude Code, Cursor, Windsurf, and Copilot. The tool exits with error code 1 to force AI agents to update documentation before proceeding.
Session Inspector for Claude Code provides real-time visibility into AI agent operations
Vibeyard, an open-source terminal IDE that wraps Claude Code, has added a Session Inspector feature that provides real-time visibility into Claude Code sessions with timeline tracking, cost breakdowns, tool analytics, and context window monitoring.