docAnalyzer.ai is a source-grounded AI document workspace for professional practitioners. According to the official introduction, its goal is to turn document overload into work worth putting your name on.
In terms of use, it targets scenarios that require extensive reading and comparison of documents. The official description points out that search can find a certain clause, but a single search cannot determine before reading what to look for, such as judging which agreements deviate from standard indemnity clauses, or comparing changes between 2023 and 2024 documents; a single search struggles to answer these. docAnalyzer.ai is positioned as a workspace rather than a chat box, emphasizing preserving context and continuing to build on it.
Its main capabilities include: three specialized chat modes - Ask docAnalyzer helps plan the next step, Focus chat runs in-depth analysis on a set of established sources, and Co-work chat optimizes and shapes output in a structured canvas editor; the three modes share the same underlying engine, and knowledge transfers among the modes. In terms of models, the official statement says 30 models from 11 providers can be used within one subscription, and users can switch mid-conversation without restarting the agent or losing context; if a model is busy, the conversation can continue with the next model.
In terms of output, the system can generate downloadable files such as PDF, spreadsheets, charts, diagrams, and packaged files. These outputs do not disappear and can later be reused, converted into new formats, or combined into packaged files. The official description emphasizes that outputs remain addressable across multiple turns of conversation: a spreadsheet generated in turn 3 can become a PDF in turn 7, become a ZIP package in turn 9, and serve as input for new analysis in turn 11.
In terms of usage, the official description says that after registering, the first session can begin within 30 seconds, with no credit card required. The free tier covers everyday document Q&A and the core workspace; the paid tiers unlock larger limits, advanced models, batch workflows beyond the included quota, and team workspaces. For the specific per-tier limits on single-file size, page count, total storage, and daily prompt usage, the official explanation is on the pricing page.
Applicable scenarios include: contract review and comparison, resume and job matching, terms and terms-of-service Q&A, summarization, large-scale document set analysis, and more. The official case mention states that Dustin's team at Rally Point Resources had 350 commercial agreements to review, each 40 to 50 pages, originally expected to take three people three months, but actually completed it within four work weeks.
Limitations and notes: supported formats include PDF, Word, PowerPoint, OpenDocument (ODT/ODP), RTF, EPUB, Excel spreadsheets, Markdown, plain text, and HTML. CSV and JSON require experimental mode to be enabled. Web pages can also be brought in via URL, and scanned pages are automatically OCRed. Standalone image, audio, and video files are not currently supported, but images within documents are automatically read. The product interface is currently in English, while documents and conversations can support multiple languages, and OCR for scanned documents supports more than 40 languages out of the box. The company states that customer content is never used to train models, workspaces are isolated by tenant, outputs are limited to the session scope, and nothing is aggregated across users. The system is not limited to the LLM context window and works through multiple rounds of retrieval, reading, and more refined searching, while long sessions remain coherent through deterministic history compression.
