
In Progress
Posted
Paid on delivery
## 项目简介 开发一款面向科研人员的本地桌面工具,用于批量读取 PDF 学术论文,自动提取论文固定章节信息,输出结构化表格。 逻辑:简单论文本地文本解析 + 固定提示词批量抽取;**识别困难、排版乱、扫描类文献,自动调用大模型 API 兜底处理**。 项目分阶段开发,第一阶段交付 MVP 核心功能,后续迭代全部高级功能,需要开发者能承接完整最终版本的开发。 ## 完整全部需求(最高需求,写在这里,让对方知道最终能做到什么) 1. 批量导入多篇本地 PDF,批量加载、排队处理 2. 自动提取论文固定字段:研究背景、研究目的、研究方法、试验材料、研究结论、研究不足 3. 智能分流机制: - 正常可解析 PDF:本地提取文本,内置提示词批量抽取信息 - 排版错乱、公式多、扫描件、识别失败的文献:自动调用 LLM API 做兜底提取 4. 提取结果可视化预览,支持手动编辑、修正 AI 识别错误 5. 全部文献结果汇总,一键导出 Excel 表格,方便横向对比多篇论文 6. 结果本地保存,可二次打开查看历史提取任务 7. 简单用户界面:文件列表、任务进度、结果面板、API 密钥配置入口 8. 后续迭代规划:账号、授权、付费版本功能(二期开发) ## 一期 MVP(第一阶段,优先开发,本次报价的范围) 只做核心基础功能,先落地最小可用版本: 1. 批量导入 PDF,本地提取文本 2. 调用提示词提取 6 个固定字段(背景 / 目的 / 方法 / 材料 / 结论 / 不足) 3. 手动配置 LLM API 密钥,识别失败时,支持手动 / 自动调用 API 兜底 4. 结果预览、手动修改,导出 Excel > > 一期**不做**:会员、复杂权限、高级图表识别、Mac 端 ## 二期迭代(一期做完之后,再协商预算开发) 完善智能分流、任务历史存储、UI 优化等高级功能,走向完整版。 ## 技术、交付、合同要求 1. 技术:可基于现有 PDF 解析开源项目二次开发(Python/Tauri/PyQt 均可) 2. 文件安全:原始 PDF 优先本地读取,不强制上传文件 3. 交付物:一期交付 exe 安装包 + 完整源码 + 使用文档 4. 约定:**开发阶段产生的 Bug 免费修复;新增功能、二期迭代单独谈价格** 5. 验收方式:分阶段验收,每完成一个模块确认后再继续下一阶段 6. 要求:开发者有 PDF 文本提取、对接大模型 API 的开发经验,需要能承接全部最终版本开发。 ## 预算与周期 本次招标为**一期 MVP 开发**,预算几百元;完整版二期功能后续单独协商。 请投标者在报价时说明:一期工期,以及完整最高需求版本的预估方案。
Project ID: 40761859
43 proposals
Remote project
Active 1 min ago
Set your budget and timeframe
Get paid for your work
Outline your proposal
It's free to sign up and bid on jobs

Hi, Local parsing handles most papers fine, but a formula-heavy page or scanned figure trips it up, and that exact moment is where the LLM fallback has to keep the batch moving. I'll build it so the handoff happens in-flight. - AUREX PDF Reader: PDF parsing + bookmark management shipped on Android. Link: [login to view URL] - AI Resume Builder: AI-driven structured extraction using OpenAI prompt templates. Link: [login to view URL] What this covers: - Local PDF text extraction with page-level error capture per paper - LLM fallback router with per-file state machine so one bad PDF never stalls the batch - Prompt template for the 6 fields with token budget control across batched calls - Editable result panel, Excel export, API key config, task persistence End state is a folder-in, Excel-out desktop tool. Phase 1: PyQt scaffold + batch import + local extraction Phase 2: LLM fallback routing + per-file state tracking Phase 3: editable result panel + Excel export Phase 4: PyInstaller build + docs Phase 2 needs a call before I lock the design. Quick question: Do you want LLM fallback to fire automatically when local parsing fails, or do you want a manual 'send to API' button per file so researchers control cost? Best regards, Mohammad Juned
$420 USD in 15 days
0.0
0.0
43 freelancers are bidding on average $393 USD for this job

Hi we can build your PDF research extraction tool using Python and PyQt. It will batch process PDFs and extract the six required fields using prompts and an LLM API fallback. We’ll include editable results and Excel export with the complete source code and user documentation. Which LLM APIs would you prefer such as OpenAI or DeepSeek? Should PDF processing remain local unless extraction fails and an API is needed?
$350 USD in 7 days
5.7
5.7

Hello, I am Dr. Rajesh Rolen, PhD in Computer Science & Engineering, with experience of over 20+ years in API Integration, Python As a preferred freelancer in the top 1%, I have done 400+ projects here on freelancer.com, I have 4.9 ratings out of 5 on average, which showcases my quality of work and timely delivery. Key Highlights: - Free Hosting Support on the Cloud or any desired platform. - Free 3 months of post-delivery support to ensure that our client doesn’t face any challenges after the launch of the project. - Free Dedicated tester on projects to ensure quality delivery, so clients don’t need to act as a tester. - 10+ Years experience UI/UX team to ensure intuitive UI. Portfolio: https://www.freelancer.com/u/Microlent Please open the chat and send me a message, so we can have a more detailed discussion about the project to give you the project timeline and cost. Thank you for considering my services. I look forward to engaging in a productive conversation and understanding how I can be of assistance in bringing your project to life. Regards Rajesh Rolen
$500 USD in 10 days
4.9
4.9

At TechWagger, we understand the importance of efficient and accurate data extraction for academic research. Our experienced team of developers, specifically skilled in Python and API integration, are well-equipped to develop the PDF Academic Article Data Extraction tool you require. We've successfully implemented similar projects before, providing clients with a clean, streamlined interface for easy importing, organizing and extracting key information from bulk PDFs. Our commitment doesn't stop at delivery. We offer post-implementation support such as free bug fixes for any issues discovered during development and look forward to collaborating with you beyond the MVP phase to incorporate advanced features like intelligent routing, tasks history storage and UI enhancement. With TechWagger, you're choosing a dedicated partner who values your project's security, aligns with its objectives and will always finish the job on time. Choose us for a robust solution that simplifies your research process from start to finish!
$700 USD in 7 days
4.4
4.4

Hello, Your MVP requires a practical local desktop tool that batch-processes academic PDFs, extracts six fixed research fields, and provides editable results with Excel export. The key challenge is making local parsing reliable while preserving an LLM fallback for difficult or scanned documents. I would build the first version with Python and a suitable desktop UI framework, keeping PDF files local by default. The workflow will include batch import, queued processing, configurable prompts and API keys, structured result editing, failure handling, and Excel generation. The codebase will remain modular for later history, smarter routing, accounts, and paid features. Similar work includes Covertly AI, where I built AI-powered workflows with React, Node.js, and API integrations, and AI content platforms involving prompt-based extraction and backend processing. This experience fits the required PDF and LLM integration logic. I will deliver a tested Windows EXE, complete source code, and usage documentation within 7 days for the MVP milestone, with staged acceptance and bug fixes for delivered functionality. I can also provide a practical roadmap and estimate for the complete version. Best regards, Bohdan
$700 USD in 7 days
3.9
3.9

您好, 我仔细阅读了您的需求,理解您希望分阶段开发一款 Windows 桌面工具,批量解析本地 PDF 论文,提取六项固定研究信息,并导出 Excel。 我可以先完成一期 MVP,同时为后续完整版本预留扩展空间。 技术上建议采用 Python + PySide6 + PyMuPDF,使用本地 PDF 文本解析,结合固定提示词和可配置的大模型 API。 我认为关键难点是不同论文的章节结构不统一,容易造成字段遗漏或提取错误。 我的方案是先提取文本并识别章节,再通过结构化提示词生成六个字段。遇到扫描件或解析失败的文件时,提示用户调用 API 兜底,并明确标记无法可靠识别的内容。 一期开发计划: 1. 批量导入 PDF 与文本解析 2. 六字段提取及 API 兜底 3. 结果预览、手动编辑、Excel 导出 4. 测试、EXE 打包及文档交付 预计一期工期:7–10 天。 二期预计增加自动智能分流、任务历史、本地存储、界面优化,以及后续账号和授权功能,预算另议。 开发阶段产生的 Bug 免费修复,新增功能单独报价。 想确认两个问题: • 一期是否需要支持扫描版 PDF 的 OCR? • 您希望接入哪家大模型 API? 期待进一步沟通。 Carlos
$250 USD in 7 days
3.6
3.6

您好!我有较强的 Python 开发和 AI API 集成经验,可以开发这款 PDF 论文批量提取工具,实现六大字段自动提取、LLM API 兜底、结果预览编辑及 Excel 导出。我会优先保证 MVP 核心功能、数据本地处理和代码可维护性,并交付 EXE 安装包、完整源码及使用文档,支持后续二期扩展。
$250 USD in 5 days
3.2
3.2

As a highly skilled Python developer, I bring to the table an abundance of experience in text extraction, PDF parsing and handling APIs that will be instrumental to the 验收方式:验收方式验收方式 project at hand. I have dealt with complex PDF documents from various sources such as wildly formatted articles or even scans, just like the application in question is designed for. My ability to carefully construct a local text solution coupled with an inherent knowledge of input errors makes me especially adept at extracting meaningful information despite damaged or unintelligible formatting. Across my 9+ years career I have led and contributed to big exit projects like yours, handing over products that have satisfied all end-user requirements. I am excited to work on your 预算与周期:预算与周期预算与周期 project, starting with crafting an MVP version with core features such as batch importing, local text extraction and word tip-based information extraction. Not only can you thus ensure a streamlined workflow but also do valid comparison between multiple papers through the resulting localized saved Excel file. I am confident in delivering this high standards social utility tool which will certainly make you return for more advanced functionalities as described in the 一期做完之后: 二期迭代 section of your project. As an added bonus, my team 缺点我进一步设计的方案 ISSUEFORALL 还为您解决了文档安全问题,在我们的方法中,原始PDF优先本地读取,并不强制上传文件。Finally以上优势加上我对接大模型API的经验足够给予你信心,能以AA高质量完成整个项目。 Thank you for considering me for this position. L
$500 USD in 7 days
2.0
2.0

Hello, 我最近在这里完成的一个项目是小说批量翻译工具:本地离线模式加上客户自己的 API 密钥模式,每章失败自动重试一次,仍失败就标记出来交给人工处理,不会静默跳过。您的工具我会用 Python + PyMuPDF 在本地提取文本,按章节标题定位背景、目的、方法、材料、结论和不足六个字段;文字层为空、乱码多或扫描件的论文自动转到您配置的 API 兜底,也可以对单篇手动触发。每篇结果在界面里预览并可直接修改,最后一键导出 Excel,一篇一行,方便横向对比。PDF 始终在本地读取,只有需要兜底的那一篇才会发送。交付 exe 安装包、完整源码和使用文档,交付前我会先用二三十篇不同排版的真实论文完整测一遍。 Amul
$290 USD in 3 days
1.2
1.2

Hi, I can build the Windows MVP in Python: batch PDF queue, local text extraction, six editable research fields and Excel export. A configured LLM API would handle extraction and difficult documents only with the user informed about external processing; keys would stay out of source code. I would include the installer, source, usage notes and testing against agreed sample papers. My Phase 1 estimate is USD 450 and 7 days, excluding API charges. Phase 2 can add saved history, improved routing and licensing as separate modules after the MVP; the database and provider layer would be kept separate for this. Your text mentions a few hundred yuan while the listing shows USD 250–750: please confirm the budget and preferred API. English communication is preferred.
$450 USD in 7 days
1.1
1.1

Your phase-one scope is clear: a local desktop tool that batch-loads research PDFs, pulls six fields (background, purpose, method, materials, conclusion, limitations) and exports them to Excel. I'd build it in Python with PyMuPDF for the local text pass, then a fixed prompt to pull those fields. When a paper is scanned or the layout breaks the text layer, the app calls an LLM automatically, with the key and provider set in the UI. Results land in an editable preview before export. You get a working desktop app, a sample run on your PDFs, and setup notes. I'll keep the code layered so phase two (accounts, licensing, paid tiers) drops in without a rewrite. Happy to run one paper end to end so you can check the field accuracy.
$540 USD in 6 days
0.3
0.3

Hi there, The challenge lies not in extracting text from well-structured PDFs but in handling those with complex formatting or those that are scanned. This project requires a solid approach to parsing and a robust fallback mechanism using an LLM API for difficult documents. For the MVP, I recommend focusing on batch importing PDFs and accurately extracting the specified fields, while implementing a manual API configuration for cases where automatic processing fails. This ensures that we meet the core requirements first before moving on to more advanced features. How do you envision the manual intervention process for editing extracted data? Thank you.
$500 USD in 7 days
0.0
0.0

您好, 需求我看完了,一期这个范围我能做。我的思路是用 Python + PyQt 做个桌面程序:先用 PyMuPDF 在本地把文字读出来,再用固定提示词抽那 6 个字段。遇到扫描件或者排版很乱的,就自动交给大模型兜底,DeepSeek、通义千问或者 OpenAI 的接口都能接,密钥在界面里自己填。结果可以先预览、手动改,最后一键导出 Excel。PDF 全程只在本地读,不往外传。 一期大概 7 天。我想分两步交:第一步先把导入和字段抽取做出来,您拿真实论文试一下效果;第二步再做编辑、导出和 exe 打包。最后给您 exe 安装包、完整源码和一份使用说明,开发期间的 bug 我免费修。 想先问两件事: 1. 方便发我 3-5 篇普通论文和 1-2 篇扫描版的吗?我拿来测试。 2. 描述里写的是"预算几百元",页面上显示 250-750 美元,一期预算大概是多少? 完整版的功能等一期验收后我们再细聊。 Dmytro
$300 USD in 7 days
0.0
0.0

Hi, I can build phase 1 as a Windows desktop tool (exe installer + source + guide) in 6 days: batch import -> local text extraction -> an automatic quality check sends hard or scanned papers to the LLM fallback (or force it per file) -> 6 fields as strict JSON -> editable table -> one-click Excel. Any OpenAI-compatible API, key stored locally. Python + PySide6. To build it right the first time, please confirm: 1. Which LLM must work out of the box (DeepSeek, Qwen, Kimi, OpenAI)? Scanned papers: page images to a vision model, or free local OCR first? 2. Papers in Chinese, English or both? Fields always written in Chinese? 3. Fields as short summaries (~100-300 chars) or quoted original sentences? Show page numbers for checking? 4. No clear section for a field (e.g. limitations): leave empty, write "未提及", or let the model infer? 5. Normal PDFs: no API at all (rule-based by headings), or local text sent to the API? This sets cost and accuracy. 6. Excel: one row per paper with file name, title, authors, year, journal + 6 fields? Other columns? 7. Typical batch (e.g. 50 papers x 15 pages)? Windows 10/11 only? Chinese UI? 8. Can you share 5-10 sample PDFs (incl. scanned) and the expected output for one, as the acceptance test? Bugs during development are fixed free; delivery module by module for your review. 您好,确认以上问题后,6天交付一期(exe安装包+源码+使用文档)。
$330 USD in 5 days
0.0
0.0

您好,我是 Java 全栈开发,熟悉 PDF 解析与 LLM API 集成,完全理解批量学术论文 PDF 信息提取工具的需求。 一期 MVP 可完成 PDF 批量导入文本提取、6 个论文字段抽取、API 密钥配置、结果预览编辑 + Excel 导出,交付 exe 安装包 + 源码 + 文档。 支持分阶段验收,开发期 Bug 免费修复,可提供一期工期与完整版预估报价。欢迎沟通细节。
$250 USD in 7 days
0.0
0.0

您好,我会用 Python 构建这套本地桌面工具,采用 PyQt/Tauri + PDF 文本解析 + LLM API 的分层架构,先完成批量导入、6字段智能抽取、失败兜底、人工修正和 Excel 导出,并保证原始 PDF 默认不离开本机。 核心不是简单把 PDF 丢给大模型,而是先判断文本解析质量:正常论文走本地解析和固定 Prompt,扫描件、乱码、公式复杂或提取失败的文件再自动/手动调用 LLM,这样能明显降低 API 成本,同时提高批量处理稳定性;结果采用结构化 JSON 保存,方便后续直接扩展历史任务、智能分流和账号付费体系。 我做过类似的 PDF 文本处理和 AI 信息抽取项目,通过本地解析 + API fallback 的方式处理复杂文档,减少了约 40% 的 API 调用量,并通过结构化校验降低字段缺失和格式错误。源码、EXE 安装包和使用文档都会完整交付,后续二期也可以继续承接。 我可以按模块分阶段开发并配合验收。请问您目前是否已经确定要使用哪家 LLM API?另外,是否有 5–10 篇真实论文样本用于第一阶段测试和验收?
$500 USD in 7 days
0.0
0.0

您好,我在高校从事科研,平时要读大量论文,很理解这个工具的用处。我主要用 Python 开发,可以按您的要求实现: 1、批量读取 PDF,按章节结构定位并提取背景、目的、方法、材料、结论、不足 6 个字段; 2、规则提取失败时调用大模型 API 补全,API 密钥和模型可在界面上配置; 3、结果导出 Excel,打包成 exe,同时交付源码和使用说明。 计划先用您提供的 5 到 10 篇样例论文做出可用版本给您试,满意后再完善。开发期间发现的 Bug 我负责修改,二期也可以继续合作。请问您那边常用哪家的大模型 API?论文以中文为主还是英文为主?
$300 USD in 7 days
0.0
0.0

IF YOU'RE NOT HAPPY, YOU DON'T PAY. I recently completed a similar project where I developed a tool that automatically extracted key information from academic papers, resulting in a 30% reduction in processing time for users. I understand you need a local desktop tool to read and extract structured information from PDF academic papers. I will ensure the tool is clean, efficient, and user-friendly, focusing on batch processing and seamless API integration for documents that may be difficult to parse. I focus on good planning, clean and maintainable code, clear communication, on-time delivery, and reliable long-term solutions. I am confident I can help you achieve your goals for this project. If this aligns with your project, feel free to reach out to discuss scope and pricing. WORST CASE SCENARIO YOU WALK AWAY WITH A FREE CONSULTATION Regards Ridwaan M
$350 USD in 7 days
0.0
0.0

您好,需求我看完了,一期 MVP 的范围很清楚:批量导入 PDF → 本地抽文本 → 固定提示词抽取 6 个字段(背景/目的/方法/材料/结论/不足)→ 解析失败时手动或自动调用大模型 API 兜底 → 预览、手动修改 → 导出 Excel。交付 exe 安装包、完整源码和使用文档。 技术方案:Python + PyQt6 做界面,PyMuPDF 抽文本,判断文本为空、过短或字段抽取失败就标记该文件,走 OpenAI 兼容接口兜底(支持 DeepSeek、通义等),可以选「自动调用」或「确认后再调用」。API 密钥只存在本地配置里,PDF 不上传。导出用 openpyxl,每篇论文一行。打包用 PyInstaller。 相关经验:我做过一个发票/报销助手(invoice-assistant),包含 OCR 识别、分类校验和 PDF/Excel 导出,同属文档识别到结构化表格这一类。作品可以提供:Portfolio available on request。 报价 $345,10 天,分三个阶段验收:① 第 3 天批量导入 + 抽文本 demo($105);② 第 7 天 6 字段抽取 + 编辑 + 导出 Excel($140);③ 第 10 天 API 兜底 + exe + 文档($100)。一期范围内的 Bug 免费修。二期(智能分流、历史任务、UI 优化、授权付费)等一期验收后,按实际样本再给方案和报价。 想先确认两点: 1. 正文写「预算几百元」,平台写 $250–750,一期按哪个币种和上限来定? 2. 能否先给 2–3 篇样例 PDF(最好含 1 篇扫描件或双栏),兜底优先接哪家模型? 杜俊言 / 汇点 AI
$345 USD in 10 days
0.0
0.0

hello, 这个项目我这边可以承接, 我这边的方案基于 deepseek harness 开发, 功能这块粗略看下来没有问题, 具体功能细节以及费用这块可以聊聊, 感谢关注 期待合作!
$500 USD in 10 days
0.0
0.0

您好,我可以按一期 MVP 交付这款 Windows 本地 PDF 信息提取工具。建议使用 Python + PyQt 构建桌面端,采用 PyMuPDF/pdfplumber 完成本地文本解析,并通过可配置的 LLM API 按固定结构提取研究背景、目的、方法、材料、结论和不足;同时提供结果预览、手动修正及 Excel 导出。 我的报价为 250 USD,工期 7 天,包含完整源码、Windows EXE、配置说明和使用文档。该报价按一期范围执行:优先处理可选择文本的 PDF,解析失败时调用 LLM API 兜底;复杂扫描 OCR、公式/表格及高级图表识别、账号会员系统不包含在一期,可在二期单独评估。 建议分三阶段验收:1)PDF 批量导入与本地解析;2)LLM 结构化提取与结果编辑;3)Excel 导出、EXE 打包、测试及文档。可以全程中文沟通,并按模块演示验收。完整版本后续可在一期架构上扩展任务历史、智能分流、UI 优化与授权功能。
$250 USD in 7 days
0.0
0.0

Taiyuan, China
Member since Oct 11, 2026
₹12500-37500 INR
€2-6 EUR / hour
$1500-3000 AUD
$30-250 USD
min $100000 USD
$30-250 USD
$250-750 USD
₹150000-250000 INR
₹1500-12500 INR
$250-750 USD
₹1500-12500 INR
₹1250-2500 INR / hour
₹750-1250 INR / hour
₹1500-12500 INR
$30-250 CAD
₹1500-12500 INR
£20000-50000 GBP
$8-15 USD / hour
€500-1000 EUR
$2-8 USD / hour