Friendli Suite

What do you like best about the product?

What stands out most to me is the impressive inference speed and consistently low latency. Model setup is quick and straightforward, and the serving engine can handle high concurrency without sacrificing performance. Overall, it makes running open-source LLMs in production faster and more cost-effective.

What do you dislike about the product?

Overall, there isn’t much to dislike, but I do wish the documentation and advanced customization guides went a bit deeper—especially for niche edge cases. A few additional step-by-step tutorials for specific custom model setups would reduce the trial and error during initial integration and make the onboarding process smoother.

What problems is the product solving and how is that benefiting you?

Serving open-source generative AI models at scale can get expensive very quickly because of GPU requirements. Friendli AI addresses this by maximizing throughput and optimizing memory usage, which significantly reduces infrastructure costs. For us, that translates into consistently fast response times for users without blowing through our hardware budget.

What do you like best about the product?

Frendli Ai offers a diverse selection of AI models. It makes API key management easy, and it’s pretty affordable, although the cost depends on the model you choose. The interface is clean and, overall, the platform is straightforward to use.

I haven’t needed to contact support because getting started was simple, and everything has performed fine so far. It also lets me integrate the AI models easily into my business workflow. Frendli doesn’t have AI built into the platform itself in the sense of a standalone AI tool, but it does provide access to the models, as I mentioned earlier.

What do you dislike about the product?

I don’t like the way they organize AI models. Sometimes it’s hard to find the specific one I used last time if I can’t remember its name. Adding members is also difficult. At times, the person I’m trying to add doesn’t receive the invite email, which is really frustrating. Because of that, I mostly use Frendli by myself and keep my members on another similar platform.

What problems is the product solving and how is that benefiting you?

Having access to so many AI models is uncommon, and with the number of models available, it basically solves any problem I run into within my workflow.

What do you like best about the product?

The fast and cost efficient LLM inference has been a real advantage for our team because we can generate and test large volumes of ad copy variations and personalized messaging at a speed that was simply not possible before and I really appreciate how straightforward the API is to work with because our team was able to integrate it into our existing campaign workflows without needing significant technical support and the token efficiency means we are getting more output for less cost which has had a noticeable positive effect on our AI spend across the quarter.

What do you dislike about the product?

The range of available model options could be expanded a little further particularly for teams looking to experiment with more specialized or domain specific models for niche marketing use cases and broader model support would make the platform even more versatile.

What problems is the product solving and how is that benefiting you?

Generating personalized content at scale used to be one of our biggest bottlenecks because the tools we had were either too slow or too expensive to run at the volume our campaigns required but now Friendli AI lets us produce high quality personalized messaging across multiple channels at a fraction of the previous cost and the efficiency gains have been significant enough that we have been able to scale our personalization efforts without increasing our overall AI budget.

What do you like best about the product?

As a Research Associate, I find Friendli AI’s support for multiple LLMs extremely useful. It gives me the flexibility to integrate different models into my research applications and compare their outputs more easily. I also appreciate how the endpoint configuration and the usage dashboard help keep the model-serving workflow clear and well organized.

What do you dislike about the product?

One limitation is that there’s no direct terminal access in the dedicated GPU environment, which can limit low-level debugging and deeper customization. Stronger debugging tools and better GPU-level visibility would make it easier to manage technical workflows and handle troubleshooting more conveniently.

What problems is the product solving and how is that benefiting you?

Integrating LLMs into research applications can be complicated, especially when you have to manage separate serving infrastructure. Friendli AI helps reduce that complexity by offering endpoint-based access, so I can spend less time on infrastructure setup and more time evaluating model outputs and building research applications.

What do you like best about the product?

What I like best about Friendli AI is how easy it makes deploying and serving large language models without having to manage a lot of the underlying infrastructure myself. The platform feels developer-friendly, and I especially appreciate the focus on performance and efficient inference. It makes it much easier to experiment with different models and move from development to production without spending too much time on infrastructure setup.

What do you dislike about the product?

One thing I dislike about Friendli AI is that the platform can feel a little overwhelming at first, especially if you are not already familiar with LLM deployment and inference concepts. Some parts of the documentation and setup process could be more beginner-friendly, with clearer examples and step-by-step guidance. Once you get used to it, the experience is much smoother, but there is still some room to improve the onboarding experience.

What problems is the product solving and how is that benefiting you?

Friendli AI solves a lot of the complexity around deploying and serving large language models efficiently. Instead of spending time setting up infrastructure, optimizing inference, and managing scaling, I can focus more on building and testing the actual application. This has helped me iterate faster, reduce the operational overhead of working with LLMs, and experiment with different models more easily.

What do you like best about the product?

As a research scholar, I find Friendli AI’s approach genuinely useful, particularly for experimenting with and evaluating different large language models (LLMs) without having to spend excessive time managing infrastructure. The endpoint-based setup is especially helpful for work like organizing scientific information, analyzing technical text, and exploring AI-assisted research workflows. I also value its emphasis on inference performance, since fast response times matter a lot when I’m repeatedly testing prompts throughout the research process. In addition, being able to configure endpoints and monitor usage through a dashboard makes the overall workflow much easier to handle. It helps me stay focused on the research itself and on evaluating model outputs, instead of getting bogged down in deployment details.

What do you dislike about the product?

I think it would be very helpful to have a detailed, side-by-side comparison of performance, GPU configurations, and estimated costs across the different models. That kind of overview would make it much easier to choose the right model and configuration before starting a research experiment. I’d also appreciate more comprehensive troubleshooting documentation, along with clearer ways to identify and diagnose issues related to endpoint configuration. Overall, the platform’s usefulness really comes down to how well its deployment options, performance, and pricing match the specific needs of a particular research project.

What problems is the product solving and how is that benefiting you?

As a research scholar, one of my biggest challenges is experimenting with different AI models without spending too much time on deployment and the underlying technical infrastructure. Friendli AI helps with this by making LLM deployment and inference more accessible through an endpoint-based approach. This is especially useful for tasks like working with scientific literature, analyzing technical information, and testing prompts for research-focused use cases. It also simplifies usage monitoring, so I can spend more time evaluating model outputs instead of dealing with infrastructure details. Overall, it makes AI-assisted research workflows more manageable and helps me explore different models based on my project requirements.

What do you like best about the product?

Offers the option to spinoff custom GPU type based on the selected model we are running. Billing is quite transparent as well. During the setup phase we get an exact estimate of the cost based on our selections, this enables us to make necessary modifications or downgrade in advance depending upon our projects's budget.

What do you dislike about the product?

They do not offer any direct terminal access to our assigned GPUs just the endpoints & a dashboard. The dedicated GPUs are billed for the entire period they are in 'active' state, forcing us to monitor the idle state & make our user requests without gaps.

What problems is the product solving and how is that benefiting you?

Friendli AI is not like other GPU rentals, their proprietary engine is faster than most open source tools that are available. The engine processes tokens faster eventually making this a cost saving option in the longer run

What do you like best about the product?

I like that FriendliAI focuses on speed, efficiency, and making AI models more accessible.

What do you dislike about the product?

I don't like that the platform seems more geared toward developers and businesses rather than everyday users.

What problems is the product solving and how is that benefiting you?

At CAMG, we used FriendliAI to help streamline daily workflows, improve efficiency, and reduce the time spent on repetitive tasks.

What do you like best about the product?

It’s easy to deploy and use any LLM without having to manage complex infrastructure. Most popular LLMs are supported, which makes it straightforward to get started and stay flexible. The interface makes it easier to explore available modes, configure endpoints, and monitor usage without needing extensive infrastructure knowledge. Performance, I think, is one of the key advantages. It's designed to deliver fast responses and handle demanding AI workloads efficiently, very valuable for applications that require low latency, high concurrent requests.

What do you dislike about the product?

Pricing and performance comparisons across different models should be explained more clearly, so I can make an informed decision.

What problems is the product solving and how is that benefiting you?

Friendli provides AI infrastructure that helps users run LLM models on their own. For startups and development teams, that’s especially useful because I don’t have to purchase expensive physical equipment, and I can freely experiment with different models while building AI-powered applications with lower upfront costs. The onboarding is fast and straightforward, and the API integrations reduce setup time and make deployment simple. It’s also flexible when working with different large language models, allowing developers to choose the models that best fit their applications.

What do you like best about the product?

Easy communication, No judgement and the guidance is really helpful

What do you dislike about the product?

Even though at times it tries to be helpful, sometimes it can misunderstand you which can be a bit annoying if you are on the go

What problems is the product solving and how is that benefiting you?

It benefits me by helping me work more efficiently, organize my thoughts, write professional messages, and find solutions to problems faster.