Helpin is an open-source AI agent workspace positioned to let teams and AI agents collaborate within the same platform, covering work such as customer support, documentation maintenance, sales follow-up, task planning, and code development. According to the official introduction, it enables AI agents to answer customer questions, update help documentation, follow up on sales leads, update customer records, turn meetings into tasks, and write code pending review.
[Purpose] Helpin centralizes customer conversations, meeting notes, documents, and project updates in one place, so that both teams and AI agents can use this information. It can be used to manage support, development, sales, and marketing work: shared inbox, task board, CRM, documents, and meeting notes, with AI agents assisting with replies, updates, and follow-ups within these tools.
[Main capabilities]
- Support scenarios: Manage chats and emails in a shared team inbox. The Echo agent can refer to documentation and check code, logs, and customer account information in connected tools before replying, can follow up on whether the customer still needs help, and can hand off to the team when necessary.
- Meeting notes: Record internal meetings and customer calls, obtain AI summaries and action items, and use Helpin AI to create follow-up tasks.
- Task planning and development: Organize tasks, stories, epics, and sprints, and track roadmap and goal progress; Scribe plans changes, Forge writes code, and Lens reviews code.
- Sales follow-up: Store contacts, deals, emails, and meeting notes in a centralized place; the Beacon agent helps identify purchase intent, understand factors blocking deals, and prepare follow-ups based on conversations.
- Documentation maintenance: Use internal documentation spaces to store team guidelines and processes; the Quill agent can prepare article updates based on new releases and unanswered customer questions, including new screenshots and browser recordings, for the team to review and publish.
[Agents and governance] The officially listed agents include Atlas (planning), Scribe (task planning), Forge (coding), Echo (support), Lens (code review), Beacon (CRM), Quill (documentation), and Mira (marketing). Users can choose which runs execute automatically and which require approval, view each agent's activity and results, and intervene when needed. Approval rules can be set for each agent; choose the documents, Helpin features, and connected services each agent can use; add built-in skills or bring your own skills; agents can also be triggered by scheduled periodic checks, or when a task changes stage or a pull request needs review.
[Usage and deployment] You can choose Helpin Cloud (provided by the officialHosting, including AI usage), or choose the open-source Community version to run on your own server and connect your own AI provider. The Community version is licensed under AGPL-3.0, the current version is Community 0.2 beta, deployed using Docker Compose, and an installation guide is provided; when self-deploying, the team manages installation, bears hosting and AI provider costs, data and backups are controlled by the user, and supported AI providers and models can be selected.
[Integrations and Automation] Support chat can be added to applications and customer information shared through Helpin's API and SDK; compatible agents such as Claude Code, Codex, Cursor, and Hermes can be connected through the MCP server to read conversations, create tasks, update documents, and run Helpin agents according to the selected permissions; Helpin agents can also use the tools of external MCP servers. AI agents can be started when tasks change or supported GitHub and GitLab events occur, for example to review new pull requests or handle tasks entering a specified stage. A command-line tool is also provided for installing the self-hosted version, checking service running status, and viewing logs.
[Use Cases] Suitable for teams that need unified collaboration across customer support, documentation, sales, and project development, especially organizations that want to control data through self-hosting and connect their own AI providers.
[Limitations] Official materials do not provide specific pricing, charging details beyond the free quota, platform availability, performance metrics, or a list of supported languages; this information is not currently available. The trial includes AI usage and requires no credit card, but the cost structure for long-term use is not explained.
