Open source · Community 0.2 beta

Your team. Your agents.
Your own servers.

Run support, docs, projects, CRM, meetings, and AI agents on infrastructure you control. Helpin Community is open source under AGPL-3.0, with no license fee or plan limits.

Self-host Helpin →View on GitHub →

Looking for managed hosting? Explore Helpin Cloud →

Open sourceDocker ComposeUnlimited users

No telemetry by default. Nothing leaves your servers unless you connect it.

Choose how you run Helpin

One connected product. Two ways to run it.

Same product either way. Cloud adds hosting, included AI, and support.

Product modules and hosting options
Capability	Open-source self-hostingAGPL-3.0 · 0.2 beta	Helpin CloudHosted by Helpin
Chat and email support	Included	Included
Knowledge and help center	Included	Included
Ask Agent and specialists	Included	Included
Projects, sprints, roadmaps, and objectives	Included	Included
CRM, deals, signals, and playbooks	Included	Included; automation varies by plan
Meetings	Included	Included
Coding agents	Included	Included
Automation and triggers	Included	Plan-dependent
Users	Unlimited	Per plan
AI models	Your own provider keys	Included AI; your own keys on Enterprise
Hosting and support	Your team; community support	Managed by Helpin, with support

The open-source product has no license fee. Your team covers hosting, operation, and AI provider usage. Cloud capacity and AI usage depend on your plan.

Need a commercial license instead of AGPL, deployment help, or support terms? Enterprise covers them. Compare plans → Read the configuration guide →

The work stays connected

The same agent workflows. Run by your team.

Use the Echo agent for support, the Quill agent for docs, and planning and coding agents for product work. Connect your own AI providers and choose the tools each agent can use.

Loading preview…

Answer with the earlier conversation in view.

Handle chat and email together. Give teammates and agents the history to investigate the question and hand it over without starting again.

Explore Support

Tilted Helpin CLI terminal showing installation checks, diagnostics, status, and logs.

From the terminal

Set it up. See how it’s running.

Install Helpin, check its health, read service logs, and run backups, restores, and upgrades from one command-line tool.

Read the CLI guide

$ helpin install

Get Helpin running.

Follow the installation checks for your machine or server.

$ helpin doctor

Check before you troubleshoot.

Inspect the installation and identify what needs attention.

$ helpin logs

Look into the service.

Read the logs behind an issue instead of guessing what happened.

From evaluation to everyday use

Try it locally. Move it to your server.

Evaluate on your laptop, then move to a server with your own domains. The CLI generates the Caddy configuration for the public setup.

Local evaluation

Start with a customer story you can test.

Use sample records to explore a conversation, a linked task, and an agent’s answer. Check whether the history helps the next person understand what needs to happen.

Prepare your machine.Check the requirements for the release.
Follow the installation guide.Set up the local services.
Set up your evaluation.Configure the connections needed for the workflow you want to evaluate.

Your local installation

Dashboard and widgetlocalhost:8085

Help centerlocalhost:8086

Object storagelocalhost:9005

Explore with sample data before bringing in customer records.

Suggested evaluation prompt: “What has this customer already tried, and what is the next step in the linked task?”

Connect an AI provider to try the agent’s answer.

Read the deployment guide

Your team runs Helpin

Keep the history.
Choose the connections.

Manage your records, files, and backups. Choose the services your agents connect to and check where those providers process information.

Keep a backup you can restore.

Back up records, files, configuration, and encryption keys. Test a restore before you depend on it.

Read the backup and restore guide

Choose the models behind the agents.

Connect Anthropic, OpenAI, or OpenRouter with your own keys and select models for the work. Set each agent’s tools and approval rules separately.

Configure AI connections

Connect the experience your customers use.

Configure your widget, published help center, and connected tools for the environment you run. Check access before enabling a new workflow.

Configure your deployment

Questions, answered

FAQs about Helpin’s self-hosting

Which parts of Helpin can we run ourselves?

All of it: support, docs and help center, AI agents including coding agents, projects, CRM with meetings, and automation. Only Cloud billing and hosted AI are left out.

What does a local evaluation require?

Linux or macOS on amd64 or arm64, with Docker Engine, Compose v2, Bash, OpenSSL, curl, and sha256sum or shasum. Allow 8 GiB RAM and 20 GiB free disk for evaluation; size production for your workload. Native Windows is not supported.

Where should we start the installation?

Run curl -fsSL https://helpin.ai/install.sh | bash, then helpin install. A manual bundle installation is also documented.

Can we choose our AI providers?

Yes. Self-hosted Helpin always uses your own providers. Connect Anthropic, OpenAI, or OpenRouter with your keys and choose the model for each agent. Knowledge search uses a separate OpenAI-compatible embeddings setting.

Who handles maintenance and upgrades?

Your team. The CLI includes backup, restore, and upgrade. Upgrades run only between releases that declare a tested path.

Who helps us if something breaks?

Community support is on GitHub and Discord. Enterprise adds deployment help and support terms.

Can we move from Zendesk or Intercom?

Import from Zendesk and Intercom is coming soon. See how Helpin compares →

Does server setup manage HTTPS for us?

No. It prepares a Caddy configuration; your team manages DNS, the proxy, and certificates.

How is Helpin licensed?

Application: AGPL-3.0-only. SDKs and widget: Apache-2.0. If your company can’t use AGPL, Enterprise offers a commercial license.

Are Kubernetes and fully offline deployments supported?

Docker Compose is the documented route. Kubernetes and fully air-gapped installations are not yet validated.

Read the self-hosting documentation

Run Helpin on your terms

Run your agents on your own Helpin instance.

Self-host the open-source product and connect your AI providers. Prefer a managed service? Helpin Cloud includes AI usage.

Self-host Helpin →View on GitHub →

Prefer managed hosting? Explore Helpin Cloud →

Your team. Your agents.
Your own servers.

Helpin, operated by your team.

One connected product. Two ways to run it.

The same agent workflows. Run by your team.

Answer with the earlier conversation in view.

One CLI to install and operate.

Set it up. See how it’s running.

Get Helpin running.

Check before you troubleshoot.

Look into the service.

Try it locally. Move it to your server.

Start with a customer story you can test.

Keep the history.
Choose the connections.

Keep a backup you can restore.

Choose the models behind the agents.

Connect the experience your customers use.

FAQs about Helpin’s self-hosting

Run your agents on your own Helpin instance.

Your team. Your agents.Your own servers.

One connected product. Two ways to run it.

The same agent workflows. Run by your team.

Answer with the earlier conversation in view.

One CLI to install and operate.

Set it up. See how it’s running.

Get Helpin running.

Check before you troubleshoot.

Look into the service.

Try it locally. Move it to your server.

Start with a customer story you can test.

Keep the history.Choose the connections.

Keep a backup you can restore.

Choose the models behind the agents.

Connect the experience your customers use.

FAQs about Helpin’s self-hosting

Run your agents on your own Helpin instance.

Your team. Your agents.
Your own servers.

Keep the history.
Choose the connections.