iFlytek Spark

DomesticFree + PaidQ&A

Document parsing, audio/video transcription, and document Q&A

Developer:iFLYTEK Co., Ltd.

iFlytek SparkIntroduction

iFlytek Spark is a large-scale model platform developed by iFlytek, integrating speech recognition and document understanding capabilities. It supports parsing PDF, Word, Excel files, as well as audio and video documents—ideal for handling meeting minutes, research reports, contracts, course recordings, and more. The platform can generate document summaries, extract key information, interpret tables, transcribe audio and video, and enable question-answering based on uploaded materials. Please note that all generated content is for reference only; critical information should be verified manually. A free tier is available for trial, while paid plans offer higher usage limits. Enterprises can integrate via API and deploy privately. With strong adaptability for long-form audio content, it works seamlessly across both web and client applications to unify analysis needs for documents and multimedia files.

Key features of iFlytek Spark

Key Product Features

1Read and parse documents in multiple formats including PDF, Word, and Excel
2Document summary, key information extraction, and clause organization
3Audio and video transcription, speaker diarization, and timestamping
4Ask questions based on documents or transcripts
5Supplement with relevant reference materials via online search
6Supports API integration to align with enterprise business workflows

iFlytek Spark Pricing Plans

Free Tier, Paid Plans, Billing Options, Private Deployment & Enterprise Pricing

Billing Type:Free + Paid

1Free Quota: New users receive a complimentary allocation; actual quota is subject to real-time platform announcements.
2Paid: Spark Membership, Enterprise API Resource Pack
3Billing: Subscription-based, Pay-as-you-go
4Private Deployment: Supports enterprise on-premise deployment

iFly Spark User Guide

Entry URL, Registration Requirements, Brief Operation Steps, and API / Client Support

Entry Point:xinghuo.xfyun.cn

1Entry: https://xinghuo.xfyun.cn
2Sign Up: Register and log in with phone number. Commercial use requires real-name verification.
3Steps: Log in to the platform and upload documents or audio/video files in the chat window.
4Summary, extraction, and Q&A instructions received.
5View results, supports continuous follow-up questions
6Export text content
7Client/API: Web, PC Client, Open API

iFlytek SparkParameter Overview

Information compiled from public sources; official information prevails.

Is data used for training?: Enterprise data isolation; individual users can manage training authorization
Underlying model: iFlytek Spark Series Large Language Models
Context length: Provide the large-window context version
Multi-modal support: Text, Documents, Images, Audio, Video
Search Online: Supports web search with source information displayed
File Parsing: pdf, docx, xlsx, mp3, mp4
Enterprise Edition and Deployment: Enterprise and on-premise deployment options available
Compliance Certification: Generative AI service registration completed
Domestic Availability: Direct connection within China. Log in with phone number. Pay with WeChat or Alipay.