Chaser exposes a read-only MCP server. Connect Claude, Cursor, ChatGPT, or any MCP client to your workspace and let your agent reason about at-risk renewals, the recovery queue, and closed outcomes — using the same data your team works from inside Chaser.
Read-only · OAuth per workspace · Revocable · No shared tokens · Included on every paid Chaser plan
Upcoming Stripe renewals predicted to fail — expiring cards, soft-decline risk, accounts that bounced last cycle.
Failed payments currently in recovery. Customer, amount, decline reason bucket, age, and whether a human review is needed.
Every terminal outcome of the recovery process: recovered, churned, written off. Reconciled and reason-coded.
Clients can query data. They cannot send emails, change playbooks, or take any action.
Every connection is authorized by you, against your workspace. No shared API keys.
Each connection sees only your own renewals, queue, and outcomes — never another workspace.
Remove access from your AI client at any time, or revoke from Chaser settings.
For Claude Desktop, add the following to claude_desktop_config.json and restart Claude:
{
"mcpServers": {
"chaser": {
"url": "https://chaser.cash/mcp"
}
}
}Cursor, Claude Code, ChatGPT, and any other MCP-compatible client accept the same URL. If your client asks for a transport, choose Streamable HTTP.
On first tool call your client will open Chaser's authorization screen. Approve once, and your agent is connected. Already signed in? Open the connectors tab to grab the URL and revoke anytime.
Other dunning tools send emails for you. Chaser lets your AI agent reason about your book — which renewals are about to fail, which recoveries need judgment, what worked last month and why. The agent proposes; your team disposes. That is the human-in-the-loop model, extended to the assistants your team already uses every day.