Obsei

Obsei is an open source, self hosted Voice of Customer platform that collects feedback from many channels, removes personal data, and analyzes what customers are saying. The project has existed since 2020 and is currently at a 1.0 release candidate. It pulls feedback from more than 20 sources plus plugins, including the App Store, Google Play, Zendesk, Intercom, Gong, GitHub, Bluesky, SQL databases and webhooks, and any REST or MCP source. Personal data such as names, emails, phone numbers, card numbers and IBANs, plus national IDs from Aadhaar and PAN to SSN, CPF and NHS, is redacted before anything is stored. Each item is then classified with either a small decision model that runs on a CPU or any OpenAI compatible LLM, grouped into themes across languages, and stored encrypted. Views hide themes drawn from fewer than a set number of people, and a record of an author can be erased on request. Results reach people and systems in several ways. A read only MCP server lets agents in Claude, Cursor, VS Code or custom tools ask questions and receive answers citing record ids, never author identities. Ordered routing rules send items to Slack, Jira, Linear or a warehouse based on label and confidence, and uncertain items go to a human reviewer. A hosted live demo uses synthetic feedback in eleven languages. The target user is a product, support or compliance team that cannot send customer data to a third party cloud. The software is free under the Apache 2.0 license with no open core tier, collects no telemetry, and runs on the customer's own infrastructure.