Mita AI Search

What is Mita AI search?

Metaso is an AI search and research tool developed by Shanghai Metaso Network Technology Co., Ltd. It first retrieves information from public web pages, academic materials, databases, or files uploaded by users, and then uses large language models to generate structured answers that include citations to the sources.

Unlike traditional search engines that primarily provide links, Meta places greater emphasis on the integration of \"searching, reading, organizing, and presenting\" information, and positions its product as one that offers \"results without ads, straight to the point.\"

PiTa AI Search has evolved from a simple search box to a tool that enables in-depth research, academic and database searches, file management, workflow handling, custom skills, slide creation, suggestions on what to learn, fact verification, and API access. Its strengths lie in the organization of Chinese-language materials, the filtering of paper information, and its ability to present that information in a structured manner; it should not be simply regarded as a Chinese version of some foreign AI search tool.

Core functions

• AI search across the entire web:After entering a natural language question, relevant web pages are searched, information is extracted, and a direct answer is generated; key elements are marked with citation tags to facilitate verification by referring back to the original pages.
• Three search intensities:It offers different levels of depth: concise for quick questions and answers, in-depth for integrating information from multiple sources. The research mode involves breaking down the problem, conducting iterative investigations, and presenting more comprehensive results.
• Multiple search ranges:In addition to the entire web, you can also choose academic resources, libraries, and podcasts. The search scope determines the types of materials that can be found; moreover, the quotas and response times may vary depending on the mode used.
• Academic search:Papers are filtered from professional journals and academic databases, displaying abstracts, author information, and standard reference details; it is also possible to configure databases in both Chinese and English. The scope of materials included in these databases and the access rights to the original texts do not necessarily correspond to those provided by school or institutional subscriptions.
• Library search:Semantic search is available for documents across the web, including industry reports, academic papers, and lesson materials; users can view or download those materials for which they have permission. Downloading and using such materials are still subject to the copyright rules of the original files as well as the platform’s regulations.
• Structured results:It automatically generates content outlines, mind maps, event timelines, as well as tables of characters, organizations, and events, to help users quickly understand the relationships between various elements in the information.
• Copy and Export:The search results can be copied or exported in Word or PDF format; they can also be transferred to the Memeta writing tool for further editing. The export formats may vary depending on the device and mode used.
• Documents and topic-specific knowledge bases:Upload personal files and create topics; searches, Q&A sessions, and organization of information related to those topics take place outside the public network. The size of the files and the volume of Q&A activity within a topic consume corresponding amounts of data.
• Slide:PPT content is created based on topics, search results, or available materials; points are typically deducted according to the actual number of pages generated, with the user agreement specifying that 1 point is consumed for each slide.
• What to learn:Audio explanations and PPTs suitable for learning are created around the given theme. The official pricing mechanism also relies on the number of slides to determine the amount of points required.
• Workflows and custom skills:Save repetitive tasks such as searching, organizing, and researching as reusable processes; this is suitable for industry research, competitor analysis, and data compilation in industries with fixed formats.
• Fact verification:It searches for evidence to support specific claims and provides verification results. This improves the efficiency of verification, but AI may still make mistakes in selecting and interpreting the sources.

Member benefits, points, and special quota prices

Mita AI search employs a combination of a free quota, membership subscriptions, and points or special quotas. Regular users can use certain search functions for free, and after registration they may receive points that are refreshed on a daily basis as well as long-term points.

Advanced research, documents, slides, topics, and generation capabilities consume varying amounts of resources in accordance with specific rules.

As of this verification, the main prices for in-app purchases available on the App Store in China are as follows:

• Monthly membership: 39 yuan;
• Annual membership: 228 yuan;
• 1000 points: 12.80 yuan;
• 10,000 points: 117 yuan;
• Search quota of 3000: 39 yuan;
• 1,000 points of knowledge base quota: 3.50 yuan;
• 10,000 knowledge points allowance: 34 yuan;
• 100,000 knowledge points: 320 yuan;
• In-depth interview analysis report: 29.80 yuan;
• 10,000 H3 points: 99 yuan.

The H3 points are used for the MiniMax H3 video generation service provided by Peita; they are not the same as regular search points, knowledge base credits, or membership benefits. The official website uses a rate of 4,399 yuan per 500,000 points as the benchmark for H3 top-ups, and these points are consumed depending on the length of the input material, the length of the output video, and its resolution.

The actual discount on activities may vary.

Points rules

Points refreshed on a daily basis are valid only for that day; any unused portion does not carry over to the next day. Long-term points are generally valid for 12 months from the date they are credited, and they become invalid after that period without any replacement or conversion into cash.

The remaining points displayed on a user’s account may include both daily points and long-term points, which are deducted by the system in accordance with specific rules.

The scope of use for gifts offered as part of promotions, points obtained through purchases, member benefits, and special quotas may vary.

Typically, one point is deducted for each slide; the audio explanations along with the PPTs also cost one point per slide. There are separate rules applicable to search depth, models, file size, special topic processing, and video generation – it’s not possible to determine whether a feature can be used solely based on the total points in the account.

Prices may vary on the website, in the App Store, and through other payment methods; before making a purchase, it is necessary to refer to the details regarding benefits, validity period, and refund policies shown on the settlement page.

Pyta API and Python SDK

PyTorch provides API key management and an official Python SDK. Developers can initiate searches, ask follow-up questions within existing sessions, receive results in a streaming format, create topics, recursively upload folders, and perform searches within specified topics.

The SDK can be installed using the Python package manager, and the keys are retrieved from environment variables; it is suitable for integrating search and knowledge base capabilities into internal tools, robots, or research workflows.

API keys should be stored only on the server side, and must not be included in web pages, mobile applications, or public code repositories. Whether membership benefits and points apply to API calls is determined in real time through the API console; there is no assumption of mutual compatibility.

For formal deployment, it is also necessary to handle rate limiting, retries, source display, cost monitoring, and user data authorization.

Open-source status

MetaAI’s search web pages, mobile apps, custom-developed models, search indexes, and the entire platform constitute proprietary services; they are not open-source search engines that can be deployed in a private environment. The official metaso-sdk repository is licensed under the MIT license, and what is made available are Python client interfaces and examples, rather than the search backend, model weights, or the source code of the platform.

Therefore, the field \"Is it open source?\" should be filled in as \"The product is not open source; the official Python SDK is licensed under the MIT license.\"

The open-source nature of the SDK only allows developers to use the code in accordance with the license terms; it does not mean that PetaSearch can be run offline, nor does it imply that there are no copyright restrictions on the API results or on the content of third-party websites.

Supported platforms

Mita AI Search offers access via web platforms as well as mobile apps for Android, iPhone, iPad, and WeChat, and a corresponding version is also available for Macs equipped with Apple chips. The App Store in China currently requires that iPhones run iOS 14.0 or later, and iPads run iPadOS 14.0 or later.

Macs with Apple chips require macOS 11.0 or later.

The authorities also provide browser extensions that allow AI-generated answers to be displayed next to common search results, as well as quick access to the search bar.

Guide to Using Mita AI Search

Complete a basic task.

1. Clarify the issue, time frame, location, source priority, and output format;
2. In Peita AI search, upload materials for which you have permission to use or enter search queries;
3. First, use AI search across the entire internet to establish a framework, and then supplement the evidence with three different levels of search intensity.
4. It is necessary to distinguish between factual information from the source, the author’s opinions, and AI-generated conclusions.
5. Check each item for dates, numbers, the original location, and any conflicting evidence;
6. The conclusions are manually revised, the verification time is recorded, and then they are published;

Create reusable professional workflows

1. Break down complex topics into four categories of questions: background, data, comparison, and conclusions;
2. A fixed set of research steps is established by combining AI search across the entire network, three different levels of search intensity, and various search scopes.
3. Give priority to using the official website, research papers, regulatory documents, and raw data;
4. A second person is assigned to review conclusions that are considered high-risk;
5. Save queries, evidence, versions, and unresolved issues;
6. Re-run after the data changes and update the conclusions;

Which users are it suitable for

• Students and researchers who need to search for papers, organize literature, and prepare research reviews;
• Operations, consulting, and finance professionals who need industry reports, information on competitors, as well as brand and market research;
• Creators who wish to quickly review the topic, organize the sources, and prepare an outline for the content;
• Knowledge workers who need to create specialized knowledge bases from personal documents and provide ongoing Q&A.
• Developers who need to access search, follow-up query, and specialized content capabilities via API.

Advantages and limitations

• The advantages of Mita include the ability to understand Chinese-language questions, the integration of information from various online and academic sources, the citation of answers, the creation of structured outline mind maps, as well as a complete workflow that goes from searching to exporting results.
• For ordinary web pages that are not Chinese papers, industry reports, or official accounts, it is generally easier to verify them than chat tools that rely solely on model memory.
• However, including citations does not mean that the answer is necessarily correct;
• The system may miss important information, misinterpret web pages, confuse people with the same name, or turn uncertain statements from the source into conclusions.
• Official agreements also do not guarantee the accuracy, completeness, and timeliness of the generated content or of third-party search results.
• For medical, legal, investment, policy-related matters, as well as for formal academic citations, it is necessary to consult the original sources, and when required, to verify them using official databases.
• Uploading files means handing the data over to the platform for processing; contracts, medical records, identification documents, customer information, and unpublished research data must first have their sensitive information removed and authorization confirmed.
• When releasing AI-generated content to the public, it is necessary to comply with the platform’s agreements and relevant rules, verify its authenticity, and indicate that it was created by artificial intelligence as required.

Frequently Asked Questions

Is Peta AI search free?

Basic functions can be tried out for free, and points may be awarded on a daily basis; however, features such as in-depth research, access to documents, special topics, slideshows, and videos require the use of points or specific credits.

The membership fee is 39 yuan per month or 228 yuan per year.

Can Mita search for papers?

Yes. The academic mode is designed for journal articles and academic databases; it allows users to view abstracts and reference information, and it supports the use of both Chinese and English databases.

Whether the full text can be downloaded depends on the original database and the user’s permissions.

Are the results from Peta search reliable?

Answers are provided along with their sources, which makes it easier to verify them compared to answers without sources; however, errors, omissions, or misinterpretations may still occur. For important conclusions, it is necessary to consult the original webpage or paper.

Can Peta upload its own information?

It is possible to create topics or knowledge bases and upload files, followed by searching for information and answering questions based on those files. The size of the files and the processing required consume quota or points from the knowledge base.

Is Peta AI Search open source?

The product itself is not open source. The official Python SDK is licensed under the MIT license; it can be used to call APIs, but it does not allow for the private deployment of the Metta platform.

Does the membership include an API?

It cannot be included by default. Customer memberships, points, access to the knowledge base, and API access represent different types of benefits; developers should check the current rules on the API Key and billing pages.