RunWhen

What is RunWhen?

RunWhen is a platform for AI agents that help SRE, DevOps, FinOps and platform engineering teams work through incident and change tickets. It pairs a library of reusable Skills, which are automation tools for infrastructure, logs, metrics and open-source software, with foreground agents that answer questions and background agents that run read-only checks to build a picture of the production environment. A skill-building harness sends every new Skill through several levels of AI code review before an agent can use it, and it can be reached as an MCP server from AI coding tools. Skills follow multi-vendor standards, so they also run as stand-alone CLI tools or inside agent frameworks such as Claude and LangGraph. Deployment can be hybrid SaaS or self-hosted, including air-gapped setups, with a bring-your-own LLM endpoint. Published pricing starts with a $30k one-time 30-day sprint, followed by annual ticket-based licenses from $25k per year and outcome-based contracts from $150k.

Use cases

• Triage incident tickets with AI agents
• Collect first-pass data before escalating
• Build safe-for-production automation Skills
• Ask root cause and configuration questions
• Copy pod logs into Jira or GitHub tickets
• Help developers self-serve CI/CD triage

Who it's for

• SRE teams
• DevOps engineers
• Platform engineering teams
• FinOps teams

Why use RunWhen

• Agents do a first pass across monitoring, APIs and databases, so more tickets can be handled without escalating to senior engineers.
• Skills are reviewed by AI at several levels before an agent can run them, which is meant to keep production access safe.
• Skills are not locked to the platform and can be used as CLI tools or with other agent frameworks if you move on.
• Hybrid SaaS and self-hosted options suit teams with strict security or air-gapped requirements.
• A ticket analysis can mine your work logs to find which automations are worth building first.

RunWhen features

• Foreground agents for root cause, configuration and cost questions
• Background agents that run read-only automation and learn the environment
• Skill library with 896 Skill tools in the public registry
• Skill-building harness with multiple levels of AI code review
• MCP server for building Skills from AI coding tools
• Agents run Skills on events, on cron schedules or with a human in the loop
• Workspace Chat, surfaced issues and on-demand tasks in Workspace Studio
• Hybrid SaaS, self-hosted and air-gapped deployment options
• Bring-your-own LLM endpoint
• Thumbs-up feedback to prioritize which new tools to build

RunWhen integrations

• Jira
• GitHub
• Slack
• Microsoft Teams
• Kubernetes
• AWS
• Azure
• Google Cloud
• Claude
• LangGraph
• VS Code

How to get started with RunWhen

1. Install the local runner with read-only cloud or Kubernetes credentials.
2. Let background agents run read-only automation for about a week to learn your environment.
3. Import Skills from the library and build new ones with the skill-building harness or MCP server.
4. Ask foreground agents questions or let them run Skills on events, schedules or with human approval.
5. Collect thumbs-up feedback from users to decide which tools to build next.

RunWhen limitations

• SaaS is offered only as dedicated enterprise installations; self-service sign-up is not available
• Pricing is annual or outcome-based; no monthly plan is listed
• The self-hosted license needs your own LLM endpoint, estimated at about $5k per year

RunWhen pricing

RunWhen is sold on custom quotes, so its prices aren't published.

RunWhen plans

• 3 Agents / 30 Skills / 30 Day Sprint

  $30k one time professional services

  one time

  • Forward deployed engineers embed for 30 days
  • Self-hosted RunWhen Agent runtime with 60 day license
  • Import Skills from the library
  • Build 30 new multi-vendor Skills

• 5,000 Tickets

  $25k /year license

  annual

  • Self-hosted Skill authoring and Agent runtimes
  • Covers all ticket system integrations
  • Bring your own LLM endpoint

• 10,000 Tickets, Chats

  $45k /year license

  annual

  • Everything in the 5,000 ticket license
  • Self-hosted Slack and/or Teams applications
  • Chat priced by threads

• 30,000 Tickets, Chats

  $115k /year license

  annual

  • All features of the smaller licenses

• Unlimited sessions

  Custom

  outcome-based contract

  • Value per ticket and per chat thread agreed with your team
  • Outcome-based contracts start at $150k

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

RunWhen alternatives

The main RunWhen alternatives listed on Debuly are PagerDuty, ResolveAI, Activepieces, Giselle, Vectara and Agent Swarm. The table compares pricing and platforms; order is A–Z.

1. 

   Activepieces

   Open-source automation platform for building flows, AI agents and tables across 770+ apps, in the cloud or self-hosted.

   Automation and agentsDeveloper platformsProductivity
   Open sourceVisit Activepieces (official site)
2. 

   Agent Swarm

   Self-hosted, MIT-licensed system where a lead agent hands work to Docker-isolated Claude Code, Codex and other coding workers.

   Automation and agentsCodingDeveloper platforms
   Open sourceVisit Agent Swarm (official site)
3. 

   Giselle

   No-code AI agent builder for creating and deploying multi-agent workflows with premium models, vector stores and GitHub data.

   Automation and agentsDeveloper platformsCoding
   FreemiumVisit Giselle (official site)
4. 

   PagerDuty

   On-call, alerting and incident response platform with AI agents, AIOps noise reduction and runbook automation for engineering teams.

   Automation and agentsDeveloper platformsProductivity
   FreemiumVisit PagerDuty (official site)
5. 

   ResolveAI

   AI agent for on-call and production engineering that triages alerts, investigates root causes and helps run incidents.

   Automation and agentsDeveloper platformsProductivity
   Contact salesVisit ResolveAI (official site)
6. 

   Vectara

   Agentic AI and RAG platform that ingests enterprise data and answers with grounded, traceable responses.

   Automation and agentsDeveloper platformsResearch and search
   Free trialVisit Vectara (official site)

RunWhen FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from cdn.prod.website-files.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.