Monday Morning, Postman Wants Me to Log In
The story starts on an ordinary Monday morning.
Chen is a backend developer at a startup. He arrived at his desk, coffee still hot, and opened Postman to test a payment callback endpoint. Instead of his familiar request panel, a login screen appeared. He remembered logging in last Friday, but the token had expired. The authentication requirement was annoying enough, but what made it worse was the company’s flaky network proxy. Postman’s cloud sync spun for twenty seconds, and the collection he created yesterday wouldn’t load.
He glanced at Wang, the frontend developer sitting next to him. Wang’s Postman was stuck on a sync conflict page. Both had modified the same environment variable yesterday, and the cloud didn’t know which version to trust.
This scene has become common in 2026. Postman killed free team collaboration in March this year, turning the free tier into a pure single-player mode. Want to share a workspace with colleagues? Pay up, starting at $19 per user per month. They also removed offline mode (Scratchpad) long ago. All data goes through the cloud by default, memory usage often exceeds 1GB, and opening a large collection produces a sluggish feeling that makes people nostalgic for the days when it was just a Chrome extension.
The developer community reacted directly. GitHub stars for alternative tools collectively surged after March. On Reddit and Hacker News, “time to leave Postman” became a content genre. People weren’t just complaining. They were voting with their feet.
Why More People Want to Leave
Postman’s problem isn’t just pricing. Pricing itself is understandable. Servers cost money, development teams need funding. What frustrates developers is the combined effect of several things:
Forced login means every API request, every environment variable, every test script can theoretically be synced to their servers. For developers working on banking, healthcare, or government projects, this isn’t about convenience. It’s a compliance red line. Can you imagine storing production environment tokens on someone else’s cloud?
Cloud lock-in means your data isn’t in your hands. Postman’s collection format can be exported as JSON, but collaboration, version control, and permission management all depend on cloud workspaces. Once you stop paying, those carefully organized hundreds of requests become like a bankbook locked in someone else’s safe.
Performance bloat creates daily friction. A tool that should just send a GET request and show the response now takes over ten seconds to start and consumes more than a gigabyte of memory during runtime. For developers accustomed to terminals that open instantly, this waiting itself erodes focus.
So the question isn’t “Postman became bad,” but “is there a tool that can return to that simple, fast, data-in-my-hands state?” The answer is yes, and there’s more than one.
Bruno: Make API Requests Part of Your Code Repository
If you search for Postman alternatives on GitHub, the project with the highest star count is probably Bruno. As of mid-2026, it has accumulated over 44,000 stars, more than Insomnia’s entire history combined.
Bruno’s philosophy can be summarized in one sentence: your API collection should live with your code.
Specifically, Bruno stores each request as a plain text .bru file, placed in a folder within your project directory. These files sit alongside your src/ and tests/ directories, tracked by Git, reviewed through diffs, and covered by pull requests. No login required, no sync needed, no network connection necessary. Open Bruno, select a local folder, and all your requests are right there.
Imagine Chen’s scenario with Bruno: Monday morning, he runs git pull for the latest code, and Bruno’s collection updates along with it. Wang changed an environment variable? It’s clearly visible in the Git diff. Conflict? Resolve it the Git way, exactly like resolving code conflicts.
Bruno supports REST and GraphQL, with gRPC in development. It has its own CLI (bru) that can be integrated into CI/CD pipelines for automated testing. MIT licensed, core features completely free.
Its shortcomings are equally clear: no built-in mock server, no API monitoring, no AI-assisted test generation. It’s not a platform but a pure request builder. If all you need is “send request, see result, run tests,” Bruno is nearly perfect. But if you need to simulate backend data during frontend development or schedule periodic health checks for production APIs, you’ll need to pair it with other tools.
Bruno has another easily overlooked advantage: it’s fast. No cloud sync overhead, no Electron bloat (though it is an Electron app, it’s much lighter), and startup and response speeds are noticeably better than Postman. For people who send hundreds of requests daily, this smoothness adds up.
Insomnia: The Veteran That Weathered Controversy
Insomnia’s story is more complex.
It originally appeared as a “clean” alternative to Postman, with a simpler interface, no forced cloud, and better GraphQL support. Many developers made it their primary tool between 2018 and 2022.
Then Kong acquired it.
In 2023, when Insomnia 8.0 launched, it made almost the same mistake as Postman: forced login to use the tool. The community exploded. The GitHub issue section turned into a protest zone. Large numbers of users switched to Bruno and Hoppscotch at that time.
Kong turned back in time, though. Subsequent versions added back Local Vault mode, allowing you to use local storage without logging in. Data stays on your machine, never touches the cloud. This made Insomnia a viable option again.
By 2026, Insomnia’s positioning sits between Postman and Bruno. It supports REST, GraphQL, gRPC, and WebSocket, covering more protocols than Bruno. It has a built-in mock server, Inso CLI for CI/CD, and supports Git Sync for version control. Its paid version (around $12 per user per month) provides team collaboration features, while the free version allows up to 3 users to use Git Sync, which is more generous than Postman for small teams.
If you’re a team of 3 to 5 people, need gRPC support, occasionally use a mock server, but don’t want to be held hostage by Postman’s pricing, Insomnia might be the most balanced choice. Its UI remains beautiful among these tools, and the learning curve isn’t steep.
But its open source license is Apache 2.0 (core parts), and the fact that Kong controls it makes some people perpetually wary. After all, a tool that has changed its stance once could do it again. No guarantees.
Hoppscotch: Open Your Browser and Start
If Bruno’s concept is “everything lives in the file system,” Hoppscotch’s concept is “everything starts from zero.”
Open hoppscotch.io, and without downloading anything or registering an account, you can start sending API requests. For the “I just want to quickly see what this endpoint returns” scenario, there’s no faster way.
Hoppscotch’s predecessor was called Postwoman (yes, that direct), but after the rename it took a more serious development path. It’s a fully open source MIT-licensed project supporting REST, GraphQL, WebSocket, SSE, and MQTT, with quite broad protocol coverage.
It offers three usage modes: browser PWA (data stored in browser local storage), desktop client (solves browser CORS limitations), and self-hosted deployment (data entirely on your own server). A mock server feature was added in late 2025, and API documentation generation (in testing) was launched in 2026.
For small teams, Hoppscotch’s self-hosted solution is quite attractive. You can spin up a complete collaboration environment in minutes using Docker Compose, with team members accessing through the internal network and data never leaving the company. Its team cloud version pricing is also reasonable, around $6-8 per user per month.
Hoppscotch’s biggest shortcoming is version control. Collections are stored in a database rather than the file system, so they can’t naturally pair with Git like Bruno. Additionally, its CLI (hopp) ecosystem isn’t mature enough yet, and its performance in CI/CD scenarios lags behind Bruno and Insomnia.
But if your use case is “frequently need to quickly test various endpoints, occasionally share a request collection with colleagues, don’t want to install a bunch of software locally,” Hoppscotch feels custom-tailored for you.
Apifox: The All-in-One Approach
The previous tools all take a subtractive approach, throwing away unnecessary parts from Postman and keeping the lightweight core. Apifox took the opposite path: it adds features, cramming nearly every stage of the API lifecycle into one tool.
Apifox positions itself as Postman plus Swagger plus Mock plus JMeter. You can design APIs (visually edit OpenAPI specifications), debug APIs (send test requests), mock data (frontend doesn’t need to wait for backend), and run automated tests (supports multi-threaded stress testing) all in the same interface. It can also generate API documentation with one click, share it online with frontend or testing colleagues, and changes to API documentation automatically sync to mock and test cases.
For Chinese development teams, Apifox has several practical advantages. First, fully Chinese interface and documentation, no language barrier for onboarding. Second, servers in China, stable access speed, no network concerns. Third, the SaaS version’s core features are free, and team collaboration features that Postman charges for are provided free to small and medium teams on Apifox.
Its smart mock feature deserves special mention. When you define an API’s return structure in Apifox, it automatically generates reasonable mock data based on field names and types. A field called phone generates phone number format, called email generates email format, no manual rule configuration needed. During parallel frontend and backend development, the mock data frontend colleagues receive looks as natural as real data.
Of course, Apifox has its limitations. First, it’s not open source. The code isn’t in your hands. Second, its collaboration model, like Postman, is cloud-first. Though it has offline functionality, the core value lies in online collaboration. For scenarios with extreme data security requirements (like government projects), this point needs consideration. Additionally, if you’re used to a Git-manages-everything workflow, Apifox’s collection management approach (database storage rather than file system) might not suit your taste.
But for most small and medium Chinese teams, especially scenarios where five or six people across frontend, backend, and testing need efficient collaboration, Apifox’s “out-of-the-box” experience is indeed convenient and effortless. Teams at companies like Didi and Meituan use it, which at least proves it can handle large-scale usage.
Thunder Client: For Those Who Don’t Want to Leave VS Code
The last option is a bit special. Thunder Client isn’t a standalone application. It’s a VS Code extension.
After installation, your VS Code sidebar gains a lightning bolt icon. Click it and you get a clean request panel. Enter a URL, select a method, fill in parameters, click send. No window switching, no opening anything extra.
For the “I’m writing code and need to quickly verify that the endpoint I just wrote returns correctly” scenario, Thunder Client’s cost of use is nearly zero. You don’t even need to interrupt your coding flow. Once the API works, continue writing the next line of code.
It requires no account, no login, and data stays local. It supports environment variables, collection management, and basic test scripts. As a lightweight API debugging tool, it has what it should.
Limitations are obvious: it only works in VS Code. If your team has people using JetBrains IDEs and people using Vim, Thunder Client can’t become a unified team solution. It also doesn’t support gRPC, has no mock server, and limited automated testing capabilities. It’s more like a Swiss Army knife you carry around. Good enough for daily cutting tasks, but you wouldn’t use it for woodworking.
Core Differences at a Glance
| Dimension | Bruno | Insomnia | Hoppscotch | Apifox | Thunder Client |
|---|---|---|---|---|---|
| Offline Support | Fully offline | Supported (Local Vault) | Desktop/PWA supported | Has offline features | Fully offline |
| Open Source | MIT | Apache 2.0 | MIT | Closed | Closed |
| Account Required | No | Optional | Not for self-hosted | Yes | No |
| Team Collaboration | Via Git | Built-in (paid) | Built-in (paid/self-hosted) | Built-in (free tier) | Not supported |
| Mock Server | None | Built-in | Built-in | Built-in (smart mock) | None |
| CI/CD Support | bru CLI | Inso CLI | hopp CLI | Supported | Limited |
| Pricing (Team) | Free | ~$12/user/month | ~$6-8/user/month | Core free | Free |
| Chinese Support | Community | No official | Community | Native | Minimal |
| Best For | Git power users | Full-stack dev teams | Quick testing/small teams | Chinese frontend-backend collaboration | VS Code users |
Recommendations for Different Scenarios
If you’re an independent developer or freelancer, mainly working alone and occasionally needing to debug endpoints: Bruno or Thunder Client are zero-cost, zero-friction choices. Use VS Code? Pick Thunder Client. Prefer a standalone app? Pick Bruno.
If you’re a small team of 2 to 5 people, need basic collaboration but have limited budget: Hoppscotch’s self-hosted solution or Insomnia’s free version (3-person Git Sync) are worth considering. For Chinese teams with an English interface barrier, go straight to Apifox. Core features are free and the Chinese experience is polished.
If you’re a technically driven team, managing everything in Git, and want API collections in your code review process too: Bruno is almost the only correct answer. Its .bru files can be diffed, reviewed, and rolled back, completely consistent with how you manage code.
If your team has separated frontend and backend development, needing mock data to let frontend proceed first: Apifox’s smart mock or Insomnia/Hoppscotch’s built-in mock server can save massive waiting time. Bruno requires pairing with external tools like WireMock or Prism in this regard.
If you work in finance, healthcare, government, or other industries with extreme data security sensitivity: Bruno (purely local data) or Hoppscotch self-hosted (data on your own server) are the safest solutions. Any cloud-first tool in these scenarios requires additional security evaluation.
One Last Suggestion
Tool choice has no standard answer, only the answer that fits your current scenario. The good news is that migration costs between these tools aren’t high. Bruno and Insomnia both support importing Postman collection JSON, and Apifox has one-click import functionality. You can easily spend an afternoon trying two or three to see which meshes best with your work rhythm.
The only certainty is this: the 2026 API tool ecosystem is mature enough that you no longer need to be held hostage by any single product. Data should be in your hands. Tools should serve you, not the other way around.



