Bug report to filed ticket
pdfbug-report-to-filed-ticket.pdfUse case · pdf
Spinrun for SaaS
Your product, support and billing tools, behind one gateway.
Connect GitHub, Linear, your helpdesk and Stripe once. Agents file the bug, draft the support reply and summarize the week, and every call they make passes rules you set and lands in a log you can read.
How it works
One endpoint, a key per pipeline, a sandbox for the big results
- One gateway URL
- ClaudeConnected
- CursorConnected
- n8nConnected
- Your product's agentConnected
One endpoint for every agent you run
Point Claude, Cursor, an n8n flow or the agent in your own product at the same gateway URL. Identity and rules are checked on every call, and an agent loads only the tools a request needs.
- Workspace keys
- support-botActive
- nightly-syncActive
- old-prototypeRevoked
A key for each pipeline
Give the support bot and the nightly sync their own workspace keys, each scoped to the apps it needs. Revoke one and the others keep running.
- Python runs in an isolated sandbox
- Files stay for the length of the session
- Only the summary goes back to the model
Big results stay out of the prompt
When a query returns more than a model should read, the agent works on it with Python in an isolated sandbox and reads back a short summary. The sandbox is part of Pro and Enterprise.
What agents take on
The work that falls between your tools
Each one is already written up as a recipe, a template or a department page.
Bug report to filed ticket
Read the thread, search the repo for the code path, check for duplicates, file the ticket and post the link back.
Use caseSupport replies from your docs
Read each incoming ticket, find the answer in your knowledge base, draft the reply and escalate what it can't handle.
TemplateFeedback, grouped by theme
Read through support tickets and get a weekly digest of what customers ask for most.
TemplateEngineering busywork
Prompts for triage, reviews, deploys and incidents across GitHub, Linear, Sentry and Datadog.
Department
Toolkits
The apps a SaaS team runs on
Connect each one once; every agent on the gateway can use it within the rules you set.
177 more apps connect the same way
- OAuth, no shared passwords
- Credentials encrypted at rest
Side by side
One gateway instead of wiring every agent
Your agents keep doing the work. The wiring, the keys and the record of what they did move to one place.
The job
Doing it by hand
With Spinrun
Connecting an agent
Doing it by hand: A separate integration for each agent and each tool.
With Spinrun: Point Claude, Cursor or your own agent at one gateway URL.
Handing out keys
Doing it by hand: One shared API key that every script uses.
With Spinrun: A workspace key per pipeline, scoped to its apps and revocable on its own.
Large results
Doing it by hand: Pasted into the prompt until it overflows.
With Spinrun: Worked on with Python in an isolated sandbox, on Pro and Enterprise.
Knowing what ran
Doing it by hand: Each tool's own logs, if it keeps any.
With Spinrun: One activity log with the user, the tool, the action and the outcome of every call.
Connecting an agent
Doing it by hand: A separate integration for each agent and each tool.
With Spinrun: Point Claude, Cursor or your own agent at one gateway URL.
Handing out keys
Doing it by hand: One shared API key that every script uses.
With Spinrun: A workspace key per pipeline, scoped to its apps and revocable on its own.
Large results
Doing it by hand: Pasted into the prompt until it overflows.
With Spinrun: Worked on with Python in an isolated sandbox, on Pro and Enterprise.
Knowing what ran
Doing it by hand: Each tool's own logs, if it keeps any.
With Spinrun: One activity log with the user, the tool, the action and the outcome of every call.
Who it's for
For teams whose product, tickets and revenue live in different tools
A bug report starts in the helpdesk, becomes a ticket in the tracker and ends up in a release note. Spinrun gives your agents one endpoint for all of it, with the rules decided once.
One MCP endpoint for Claude, Cursor or an agent you built
Read, write and destructive rules for each app, on Pro
A workspace key per pipeline, revocable on its own
Plan fit
Plans for agents in production
Pro is for the agent you're putting in front of users, with the sandbox, the agent database and Enhanced Control. Enterprise adds custom volume, terms and a dedicated SLA.
Pro
For the agent you're putting in front of users.
- credits / month
- 30K
- connections
- Unlimited
- team members
- Unlimited
- All of Free, and:
- Teams — give each group only the connections it needs
- Brain — indexed knowledge your agents search in one call
- Enhanced Control — read / write / destructive
- Artifacts — the agent writes and versions deliverables you keep
- Code sandbox — the agent runs Python against your tools
- Agent database — private tables per agent, shared tables per workspace
Enterprise
Your volume, your auditors, your terms.
- credits / month
- Custom
- connections
- Unlimited
- team members
- Unlimited
- All of Pro, and:
- Custom volume, terms & onboarding
- Dedicated SLA & account team
Use cases
Recipes to start from
Worth knowing before you wire it in
Yes, if it speaks MCP. Clients that support OAuth sign in as the person who approved them; anything else sends a workspace API key as a header.
Put one gateway in front of your stack.
Start free with one gateway URL, then add rules and keys as more agents come online.