As an AI search engine designed for programmers, how does Devv perform? In this article, the author has made a simple analysis of the product Devv, from the core technology, function dismantling to the profit model, let's take a look.
1. Basic Information
In a nutshell: Devv is an AI-powered search engine designed for programmers to help developers quickly find programming solutions and related resources through intelligent code search and personalized recommendations.
Logo：
Structurally:
The infrastructure is a small 3D square (data box), which symbolizes the program data block, which intuitively feels as a collection of large amounts of data. The right side of the square is three-dimensional through the reverse color, and a white bar looks like a folder pulled out, which is combined with the block as a whole, like extracting a file from a document box in an archive, suggesting its powerful search engine function
Color:
Only black and white are used, which is simple and clean, representing its concept of efficient search; The black and white colors are like the opposite state of 0 and 1 in machine language, which also implies that they are products provided to programmers
Slogan＆Mission：
Slogan: "The Next Generation AI Search Engine for Programmers Best"
Plainly expressed as a programmer for the market
Mission：“让编程更简单”
It is the core concept of Devv, emphasizing its mission to improve development efficiency through intelligent tools
2. Functional dismantling
1. Programmatic search
In addition to supporting the search function of general engines, it also supports a variety of commonly used programming languages, including but not limited to Python, Java, JavaScript, C, C#, etc., and also supports switching between different AI responsible for searching, such as GPT-4, etc., and attaches a reference website link while answering questions to help users quickly find relevant code samples and solutions.
2. Personalized recommendations
Based on the user's search history and preferences, personalized content recommendations are provided at the bottom of the search bar, including common questions, open-source tools, etc., and clicked to jump to the corresponding AI answer interface to improve learning efficiency.
3. Smart conversations
Enhance the user experience by allowing users to interact with the system and continue to ask questions about answers.
4. Personalized needs:
Customize how Devv interacts with your users by providing personal information, setting rules for AI answering questions, and choosing your preferred response language to deliver a more personalized and efficient experience that meets your users' needs.
Third, the core technology
Devv AI's core technology is based on LLMs (Large Language Models), which not only understand and generate natural language, but are also specifically trained to optimize the response quality of programmatically relevant queries. As a result, LLMs require a lot of computational resources and datasets, which happen to be provided by Google's massive data.
Devv AI also applies RAG (Retrieval Augmented Generation), which combines the advantages of information retrieval and generative models to improve the accuracy and relevance of generated content. To ensure the performance of RAG systems, Devv AI has built an automated evaluation framework that assists developers in quickly evaluating and optimizing models.
Fourth, the profit model
1. Subscription Plans
Devv AI offers two main subscription plans:
- Monthly subscription: Users pay $15 per month and enjoy unlimited searches, intelligent and accurate Agent Mode, advanced language model support, and more.
- Annual subscription: Users can choose to pay $144 once, enjoy the same service, and often have a discount (like the current 20% startup discount). Annual subscriptions are more cost-effective than monthly subscriptions, encouraging users to opt for long-term use.
This tiered pricing strategy allows Devv AI to appeal to users with diverse needs, from occasional users to developers who rely on its services for a long time.
2. One-time payment
In addition to subscription plans, Devv AI also offers a one-time payment option. Users can enjoy all the premium features by purchasing Devv Pro for a one-time payment. This model is suitable for users who don't want to pay on a regular basis, or who want a full experience before using it.
3. Upgrade and Conversion Mechanism
Devv AI allows users to flexibly switch between monthly and annual plans. For example, if a user is on a monthly plan and wants to upgrade to an annual plan, it can be done with a simple operation. This flexibility increases user stickiness and may encourage more users to switch to an annual subscription.
5. Deficiencies and iterations
- Information grading: You can mark the priority of the information source link, reducing the time for users to filter more specific information.
- Expand the diversity of answers: You can classify and display the guesses of the likelihood of the answer, instead of just answering and pushing a certain definition, otherwise users may not be able to query the desired definition. For example, when faced with an English abbreviation, a variety of possible explanations can be listed for the user to choose from.
- Natural Language Recognition Localization: The same term should have the same definition in different natural languages, but for example, ground truth (annotated data, computer language), in the case of Chinese and English, the AI search will get different interpretations. AI can be further trained to localize language expressions to better serve people in different languages.
- Strengthen the ability to recognize technical terms: It can strengthen the training of nouns in various technical directions, and reduce the inability to understand some professional vocabulary or its abbreviations, such as NRD (NVIDIA's real-time noise reduction technology), the AI answer is incomprehensible, and the explanation of guessing has nothing to do with the technical field, which is deviated from the concept of exclusive programmers.
Sixth, the strategic direction
- Integrate with more developer platforms: Collaborate with more popular developer platforms to drive traffic to other platforms and expand your customer base
- Community-based transformation: Expand the community communication function and provide a platform for users to post and communicate. After the scale is formed, the incentive mechanism can be used to form a paid essence post, from which a cut can be made: or introduce technology-oriented product advertisements to monetize community traffic
This article was originally published by @笑笑生观察日记 on Everyone is a Product Manager. Reproduction without the permission of the author is prohibited
The title image is from Unsplash and is licensed under CC0
The views in this article only represent the author's own, everyone is a product manager, and the platform only provides information storage space services