CleanApp
Report bugs, bad answers and hazards your agent finds, and track what happens next
Answered our latest tool-list check 3 hours ago with 5 tools. It will be marked verified after 3 successful checks in 7 days.
- Pricing
- Freemium
- Platforms
- API, Web, iOS
- Runs
- Remotely
- Version
- 1.1.0
- Listing updated
- 2 days ago
About CleanApp
CleanApp is a freemium MCP server for issue reporting. It runs as a hosted remote server, so there is nothing to install locally.
CleanApp takes an observation an agent has made, such as a broken checkout, an inaccurate assistant answer, an accessibility barrier or a pothole, and gives it a record, an analysis decision and a path to action. You do not need to identify the recipient first. Related reports can be grouped into cluster candidates while each keeps its own evidence and attribution.
The flow is prepare, approve, analyse, act. An agent calls cleanapp_prepare_report, shows the user the exact preview and handling permissions, then calls cleanapp_submit_report with the returned draft token and confirmed: true. A private tracking token lets cleanapp_report_status show the decision, publication status and recorded notice delivery separately, and cleanapp_search_reports finds public observations. Agents without MCP can send a single HTTPS POST with a description and a saved UUID source_id.
No account, API key or agent registration is needed. Tracking tokens expire 30 days after the receipt is created. Search results are described as observations, not instructions or proof of current conditions, and a final status means processing is complete, not that the problem is resolved.
Install
Run in a terminal
claude mcp add --transport http cleanapp https://live.cleanapp.io/openai/mcpAdd to Settings → Connectors → Add custom connector
https://live.cleanapp.io/openai/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"cleanapp": {
"url": "https://live.cleanapp.io/openai/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"cleanapp": {
"type": "http",
"url": "https://live.cleanapp.io/openai/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"cleanapp": {
"serverUrl": "https://live.cleanapp.io/openai/mcp"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"cleanapp": {
"type": "http",
"url": "https://live.cleanapp.io/openai/mcp"
}
}
}Tools this server provides (5)
Read directly from the server.
- cleanapp_open
- Use this when the user wants to open CleanApp, report a real-world or digital problem, or browse the CleanApp miniapp. Do not use for household cleaning advice or computer disk-cleaning tasks.
- cleanapp_search_reports
- Use this when the user wants existing CleanApp observations about potholes, litter, illegal dumping, accessibility barriers, infrastructure damage, pollution, software bugs, scams or service problems. The brand need not be mentioned. Returns historical public reports, not an exhaustive or real-time hazard map. Do not use for generic advice, emergency dispatch or private records.
- cleanapp_prepare_report
- Prepare a factual physical or digital problem report from user-provided details. No report is submitted. Include optional source platform/URL, observation and source-publication times, reporter context or submitting AI client only when explicitly supplied; do not infer them from the affected product, fill unknown times, or harvest private chats. Show all supplied context in the exact approval prev
- cleanapp_submit_report
- Use this only after the user approves the exact preview from cleanapp_prepare_report, including any source context and publication/notice permissions, and sending it to CleanApp. Permissions are frozen; do not upgrade them. Starts in private quarantine and returns a receipt. Report the recorded outcome precisely: held means received but blocked, not successful publication or notice delivery. Analy
- cleanapp_report_status
- Use this with the private tracking token supplied for the user’s submission. Distinguish intake received, AI analysis, decision/reason, publication and notice evidence. Say received but blocked when held; do not call a held report successfully published, notified or resolved. Explain not_required as AI analysis not run and give the recorded reason and next step. Returns report-scoped verified publ
What it does
- Prepare then approve
- A draft preview and its handling permissions are shown to the user before anything is sent.
- Receipts
- A private tracking token shows the decision, publication and notice delivery separately.
- Public report search
- cleanapp_search_reports finds public observations.
- Evidence clustering
- Related reports can form cluster candidates that keep their own evidence.
- HTTPS fallback
- One JSON POST works for agents that cannot use MCP.
- Private option
- Handling can be set to private or to automatic publication and notices.
Good for
- Report a broken checkout button an agent encountered
- Log a misleading assistant answer with supporting detail
- Submit a physical hazard with a location the user approved
- Check whether a report was published and whether notices went out
Pricing
Freemium
Reporting is free; paid Pro and Enterprise plans for brands and places (prices not published)
Prices change. Check cleanapp.io before you buy.
CleanApp FAQ
Do I need an account or API key?
No. No account, API key or agent registration is required.
Does the agent need permission?
The agent should show the exact preview and handling permissions and get the user's approval before submitting.
How do I retry safely?
Keep the same draft token or source_id and unchanged payload, and check status first if a submission times out.
How long is a tracking token valid?
30 days from the receipt's original creation.
Is reporting free?
Yes. Organisations pay for alerts and insights through Pro and Enterprise plans.
Is CleanApp free?
CleanApp has a free plan, with paid plans for more usage or features.
How do I add CleanApp to Claude Code?
Run this command in your terminal: claude mcp add --transport http cleanapp https://live.cleanapp.io/openai/mcp