Your team finally standardized on HarborClient for API work and then someone asks, “Where’s the latest collection?” Sharing shouldn’t mean emailing exports, chasing someone’s laptop, or handing out database passwords to every new hire.
HarborClient Team Hub is the answer for teams that want a central place for shared API collections without the overhead of shared database credentials on every machine. You run the server on infrastructure you control. Teammates connect from HarborClient with a URL and their own access token. That’s it.
What is HarborClient Team Hub?
Team Hub is a self-hosted Linux server and command-line tool that works alongside HarborClient. It is not a replacement for the desktop app, it is the backend your team connects to when you want everyone working from the same shared collections.
On the server, Team Hub stores the things teams actually need to share:
- Collections, folders, and saved requests
- Collection-level variables, headers, and auth defaults
- Pre-request and post-request scripts
HarborClient desktop clients talk to Team Hub over a straightforward HTTP API. Your ops person runs one server; developers keep using the app they already know.
How it works for your team
You do not need a deep technical background to understand the workflow. Here is what rollout typically looks like:
- Your ops person deploys Team Hub on a Linux server your team can reach an office VPN, a cloud VM, or a dedicated internal host.
- They configure storage using Postgres, MySQL, or Firestore, plus Redis for secure login throttling.
- They create user accounts and API tokens. Each teammate receives a bearer token (prefixed with hbk_) scoped to the collections they should see.
- Teammates connect in HarborClient via File → Team Hubs, enter a display name, the hub URL, and their token. Shared collections appear in the sidebar.
In plain terms: HarborClient on each laptop → Team Hub over HTTPS → shared collections in your database.
For step-by-step instructions on the HarborClient side, see the Team Hubs guide in HarborClient: https://harborclient.com/team-hubs
Why team leads like it
If you are deciding how your team should share API work, Team Hub is built around a few principles that matter to people who run teams — not just people who run servers.
- No database credentials on every laptop. Teammates need only the hub URL and their personal token. You are not distributing connection strings or database passwords across the org.
- Fine-grained access control. Grant access to all collections, specific collections, or narrow scopes per user. When someone leaves the team, revoke their token — you do not rotate shared infrastructure passwords for everyone.
- Separate admin and user roles. Ops accounts manage users and tokens. Developer accounts work with shared collection data. Clean separation of duties.
- Self-hosted, your rules. Collection data lives on infrastructure you choose. Team Hub is MIT-licensed and designed as a companion to HarborClient — no vendor lock-in, no surprise cloud bills.
- Optional centralized AI. If your organization wants AI-assisted API work, Team Hub can proxy requests to OpenAI, Claude, and Gemini using keys stored on the server — not pasted into every developer’s settings. Admins can enable access and set monthly token limits per user.
- Built-in protection against brute-force attacks. Failed authentication attempts are throttled via Redis, so a leaked or guessed token does not get hammered indefinitely.
Team Hub vs. other sharing options
HarborClient supports several ways to work alone or together. Team Hub fits a specific niche: ongoing team sharing with simple, token-based access.
| Approach | Best for |
|---|---|
| Local SQLite | Solo work on one machine |
| Shared remote database | Teams comfortable sharing database connection details |
| Encrypted invites | One-off handoff of a single collection |
| Team Hub | Ongoing team sharing with token-based access and no per-person database setup |
One important expectation to set upfront: environment variable groups stay local on each developer’s machine. Team Hub shares collections — the requests, folders, and collection defaults your team builds together. Per-developer environment files (your local “Production” or “Staging” overrides) are not synced through the hub today. That keeps personal configuration personal while shared API definitions stay central.
What lives on the server
When a collection is stored on Team Hub, your team shares a single source of truth for that API workspace. Folders organize saved requests. Collection-level variables, headers, auth settings, and scripts travel with the collection. When someone updates a request, the change is written to the hub — and teammates see it the next time HarborClient reloads data from that connection.
Team Hub’s API also supports server-side environments, but HarborClient currently keeps environment groups in each user’s local registry. If shared environment sets across the whole team are on your roadmap, that is worth knowing — but today, plan on collections as the primary shared asset.
Getting started (hand this to your ops person)
Setting up Team Hub is a one-time infrastructure task. The project ships as a Node.js CLI (team-hub) with full documentation at https://harborclient.github.io/team-hub/
For local development, docker compose up starts Postgres and Redis. Copy server.yaml.example to server.yaml, run team-hub migrate to apply the database schema, then team-hub start to bring the HTTP server online. Create users and issue tokens with the CLI — your teammates never need shell access to the server itself.
The source code and release notes live on GitHub: https://github.com/harborclient/team-hub
Detailed setup, authentication, and CLI reference:
- Setup: https://harborclient.github.io/team-hub/setup
- Authentication: https://harborclient.github.io/team-hub/auth
Honest limitations
Team Hub is deliberately simple — and that comes with straightforward tradeoffs worth knowing before you roll it out:
- No background live sync. Changes from teammates appear when HarborClient reloads hub data — for example, after restarting the app or re-saving the hub connection. There is no constant polling in the background.
- Concurrent edits use last-write-wins. Two people editing the same request at the same time will not get automatic merge conflict resolution.
- If the hub goes offline, sidebar entries for hub-backed collections remain, but you may see warnings until connectivity returns.
For most teams sharing API collections, these are reasonable compromises for a self-hosted model that stays easy to operate.
Ready to connect your team?
If you are a team lead evaluating how to share HarborClient work, Team Hub gives you a clear path: one server, individual tokens, shared collections, and data on infrastructure you control.
- Ops and admins: Read the Team Hub documentation (https://harborclient.github.io/team-hub/) and explore the GitHub repository (https://github.com/harborclient/team-hub).
- Developers: Download HarborClient, ask your admin for a hub URL and token, and open File → Team Hubs.
- Questions or feedback: Open an issue on GitHub or follow along as the project evolves.
HarborClient was built for developers who want a capable API client without accounts, subscriptions, or lock-in. Team Hub extends that same philosophy to teams: your collections, your server, your rules.
Leave a Reply