Open Comet

• 1. overview (780 chars)

  Open Comet is a Chrome browser extension created by Indian independent developer Prince Chouhan. Its main purpose is to "let AI control the browser like a real person and complete in-depth research tasks autonomously." The product launched on Product Hunt on April 13, 2026, ranking 11th that day, with 102 votes and 111 followers. The core selling points are zero data upload, local screenshot processing, privacy-friendly architecture that supports BYOK, and compatibility with more than 6 AI vendors including OpenAI, Anthropic, Gemini, Groq, Mistral, and Ollama.   The product is currently in its early stages, with only 5 stars on GitHub and no coverage in mainstream tech media yet. For technical users who are willing to try new things, this is a prototype of an automation tool worth paying attention to; but for ordinary users who want a stable and reliable product experience, the maturity at this stage may not be enough.

• 2. Product background (940 chars)

  The track of browser AI automation has continued to gain popularity in the past two years. From Anthropic's Computer Use to Microsoft's Copilot+ to various open source projects, they are all trying to let AI take over repetitive web page operations. Open Comet's entry point is slightly different - it is not an independent desktop client, but exists directly as a Chrome extension, which minimizes the barriers to use.   The developer Prince Chouhan (GitHub account princechouhan19) is currently an independent developer with no team endorsement and no institutional financing information disclosed. The product is open source under the MIT license, the latest version is v1.1.0, and the code is hosted on GitHub. Judging from the submission record and version rhythm, this is an early project with active iterations, and features are still evolving rapidly.   It is worth mentioning that the name "Comet" is somewhat confusing in the market. Perplexity also has an AI browser called Comet, which is completely unrelated. Prince Chouhan made this clear in a Product Hunt Q&A, saying the naming was purely coincidental.

• 3. Product functions and usage experience (1863 chars)

  Open Comet's core capabilities revolve around the three dimensions of "autonomous browsing + in-depth research + structured extraction".   Independent browsing is the basis. The extension can receive natural language instructions, then open the web page in the browser, click on elements, fill in forms, and scroll the page. The entire process simulates real user operations. This capability is supported by the "self-healing runtime" - when encountering page structure changes or unexpected errors, the system can automatically identify and adjust the execution path instead of directly reporting an error and stopping.   In-depth research is one of the highlight features, inspired by Stanford's STORM paper (Synthesis of Topic Outlines through Retrieval and Multi-perspective question asking). The user proposes a research topic, and the extension will break down the problem, retrieve information in multiple rounds, and comprehensively generate a structured report. The developer frankly stated in the Q&A that this is closer to "iterative search integration" than academic-level citation tracking, and is not suitable for formal academic writing.   Data extraction supports exporting web content to JSON or CSV format for further analysis and processing.   In terms of human-machine collaboration, Open Comet has introduced an "approval guard" mechanism - before performing sensitive operations, the system will pause and request user confirmation to avoid the risk of misoperations caused by AI autonomous decision-making. This design is not common among automation tools and reflects developers’ thinking about the “sense of AI boundaries.”   There is also a "skills system" that allows users to pre-define the execution process of common tasks, similar to macro recording, for easy reuse.   Privacy by design deserves its own emphasis. All screenshots are processed locally and are not uploaded to any server; BYOK is supported, users bring their own API Key, and developers do not handle any API call data; it also supports running local models in a completely offline state through Ollama. This architecture is obviously attractive to data-sensitive business users or individual users.

• 4. Pricing and business model (768 chars)

  There are currently two levels of pricing:   Student (free): Basic functions are available, suitable for individuals to try out.   Premium (₹299/month): Approximately RMB 25, equivalent to approximately US$3.5 in US dollars, the pricing is quite affordable. For users who need more frequent or advanced functionality.   The product adopts the MIT open source license, and the code is completely open, which means that technical users can deploy it by themselves without paying. The business model is currently very early, and there is no clear information on the commercialization path of the enterprise version or API yet.   Compared with similar products (such as Browser Use, Skyvern, etc.), Open Comet's pricing strategy is obviously more geared towards developers and student groups. The short-term profit expectations are not high, and it is more like it is in the stage of accumulating users and word of mouth.

• 5. User reviews (962 chars)

  Since the product has just been launched, real user feedback is still relatively limited, mainly concentrated in the Product Hunt comment area.   Supporters generally approve of its privacy architecture and open source transparency. Some users said, "The combination of BYOK + local processing makes me willing to give this type of tool a chance." Some developer users commented, "The design idea of ​​the self-healing runtime is interesting, which shows that the developers have considered fault tolerance issues in actual use."   Criticisms focus on feature maturity. Some people pointed out that "the depth of operations that Chrome extensions can do is limited, and they may get stuck when encountering complex dynamic pages." Others expressed concerns about "GitHub only has 5 stars", believing that the community support is insufficient and there is uncertainty in subsequent maintenance.   The developer's response in the Q&A was quite candid, and he did not shy away from the limitations of his product (the inability to bypass bot detection, limited support for academic citations), which is rare among developers of early products.

• 6. Industry evaluation (783 chars)

  There is currently no coverage from mainstream tech media—TechCrunch, Wired, The Verge, Hacker News, etc. This is not only because the product is too new (it was launched on the same day), but also reflects its gap in industry visibility. There is also a blank in Chinese technology media. A few AI tool aggregation platforms include it but do not provide in-depth reviews.   From an industry perspective, the competition in the "browser AI agent" track where Open Comet is located is quite fierce. There is greater resource investment in Anthropic Computer Use, OpenAI Operator, Skyvern, Browser Use and other directions. As an individual developer project, Open Comet is difficult to compete head-on with these opponents in terms of technology accumulation and engineering quality. Its survival space lies in whether the differentiated positioning of "lightweight, privacy, and low cost" can form a stable segmented user group.

• 7. Controversies and Risks (923 chars)

  The risk of single-person maintenance is the main hidden danger. The entire project is currently maintained by Prince Chouhan alone. Once the developer stops updating, the product will quickly become outdated, especially in the context of frequent changes to the Chrome extension API.   Blurred functional boundaries are also potential problems. "Autonomous browser control" tools will encounter a large number of edge situations in actual use: login pages with verification codes, dynamic loading of content, anti-crawling mechanisms, etc. The developers have admitted that there is no way to bypass bot detection, which limits the practical use cases.   The two-sided nature of the MIT protocol: open source means transparency and verifiability, but it also means no commercial constraints - if someone makes a fork based on this code that abuses privacy, the original developer has no control.   Although the naming confusion has been actively clarified by the developers, the confusion with Perplexity Comet will continue to affect user perception at the search and communication levels.

• 8. Applicable groups and usage suggestions (860 chars)

  Open Comet is best suited for the following types of users:   Independent developers and technical users who are keen to try new things, especially those who have browser automation needs but do not want to pay high prices for enterprise-level tools. The open source architecture allows deep customization, and the MIT license also facilitates integration into your own projects.   For researchers or freelancers with privacy concerns, and those who need AI-assisted information collection but are unwilling to upload data to the cloud, the combination of BYOK + local processing is a reasonable choice.   For students, the free version is sufficient for light use, and the payment threshold of ₹299/month is extremely low.   Scenarios that are not recommended: automation of enterprise core business processes, production environments that require high stability, complex tasks that require high automation depth (such as cross-platform account operations, scenarios that require bypassing security verification).

No in-depth report yet

• 

  Bobby_CarterJr419

  —

  刚装上 Open Comet，第一印象不错。隐私这块确实做到位了，截图全本地处理，不往服务器传数据，这点很戳我。需要自己的 API key，稍微有点门槛，但考虑到完全免费+开源，值

---

• 

  PamelaChavez_9943

  —

  试了一下研究功能，同时跑几十个来源确实快，比我手动搜强多了。引用这块做得比较扎实，不像某些工具瞎编

---

• 

  Jeffrey.Thomas_66

  —

  数据提取导出 JSON/CSV 很好用，做竞品分析的时候直接抓表格数据，省了不少功夫

---

• 

  angrysnake105

  —

  用 Ollama 跑本地模式，速度还行，关键是数据不出本机，安全感拉满

---

• 

  DrPaulineWilliams_x

  —

  BYOK 好评，能接自己的 API key，不用被绑定

---

• 

  Megan_Bennett007

  —

  Chrome 扩展形式好评，不用换浏览器，直接在现有流程里用，切换成本低

---

• 

  Roger_Taylor

  —

  技能系统有意思，能把常用操作存成宏，后续调用方便

---

• 

  Denise_Rogers

  —

  学生计划免费一天体验，薅到了

---

• 

  iThomasLaufer_pro

  —

  支持 6 家 AI 提供商，OpenAI、Claude、Gemini 都能用，灵活

---

• 

  PeterJones_2022

  —

  说是免费但得自己提供 API key，无语了，普通用户用不了吧

---

• 

  Eric.Peterson_Max

  —

  高级版 299 卢比/月，印度定价？国内用着有点怪

---

• 

  Lauren.EvansQ

  —

  安装挺方便的，Chrome/Brave/Edge 都能装

---

• 

  Diane.White_77

  —

  开源项目 v1.1.0，透明度这块比很多闭源 AI 工具强

---

• 

  JWard369

  —

  研究循环确实好用，交叉引用信息这点比单纯搜一下强

---

• 

  兰花279

  —

  零服务器数据存储，隐私优先，这个理念我认同

---

• 

  Melissa.Murphy29

  —

  用了一段时间，感觉挺稳的，没出过什么幺蛾子

---

• 

  Roger_Williams520

  —

  和 Perplexity 的 Comet 不一样，这个是 Chrome 扩展，更轻量

---

• 

  Kimberly.Lopez_Max

  —

  自我修复机制有点东西，任务跑偏了会自动纠正

---

• 

  DanielleReed_2022

  —

  置信度评分有用，至少知道它自己判断准不准

---

• 

  redlion584

  —

  高风险操作要审批，这个设计很合理，不用担心误操作

---

• 

  MRichardsonX

  —

  目前还是 beta 阶段，功能在快速迭代中，期待后续更新

---

• 

  brownkoala301

  —

  唯一问题是只支持 Chromium 内核，Firefox 和 Safari 用户没法用

---

• 

  Nathan_ScottZ

  —

  需要注册账户还要 JWT 认证，便利性打了点折扣

---

• 

  Jason.Brooks_X099

  —

  本地截图处理这个很戳，数据安全意识很强

---

• 

  Mark_ParkerJr

  —

  对比了下其他 AI 浏览器，Open Comet 的隐私保护是最严格的

---

• 

  董艳

  —

  研究效率确实提升了，尤其是多源整合这块

---

• 

  RachelLopez_20232

  —

  希望后续能出个 Safari 版本，Mac 用户表示很需要

---

• 

  Sharon.Alvarez168

  —

  用了两个月，自动化任务这块确实香

---

• 

  EDiaz_2024876

  —

  对于技术用户来说很友好，非技术的话还是有点门槛

---

• 

  Harold.Vasquez_Plus

  —

  结构化报告输出，研究用很方便

No reviews yet