Frequently Asked Questions
Free or unlimited plans often come with hidden restrictions. These services typically limit context length, response size, and request frequency. DocXee provides transparent pricing with no hidden limitations, giving you the full power of each AI model.
DocXee provides access to the full features of each AI with maximum context length and maximum outputs. We operate on a pay-as-you-go model without requiring monthly commitments, giving you flexibility and full control over your AI usage.
A system prompt is an instruction set that guides AI behavior and sets operational parameters. Most providers restrict system prompt modifications to enforce their terms of service and prevent misuse. DocXee allows customization of your system prompts, giving you maximum flexibility for your specific needs.
Not necessarily. While powerful, DeepSeek R1's extended thinking process is often unnecessary for most tasks and the improvements over our recommendation (Claude 3.7) are marginal. The extra processing time is hard to justify.
Prompt Library is our curated collection of system prompts created by prompt engineers. They make it easier to perform common tasks with AI, like translation or paraphrasing.
AI models sometimes make subtle, unintended changes to text such as altering spelling, grammar, or punctuation. These modifications may reveal use of AI, create inconsistency with your own style or break computer code. To highlight what was changed, we provide a Track Changes options in AI conversation and we provide a separate Track Changes copy for file modification.
DocXee has the best value-for-money compared with most other providers. As an example, suppose OpenAI charges us $0.10 for your question, we'll only charge you ~$0.25 after subscriber discounts. Unlike competitors with misleading 'unlimited' plans, we provide clear information on per-question and per-file charges, viewable in your 'Account' page.
Most providers arbitrarily cut off the previous parts of your conversation to save costs. They often don't tell you what has been removed. That's why sometimes AI forget your first question or cannot view the sources you quote. In DocXee, you can be confident that every part of your conversation is sent to AI, regardless of costs
a. When appropriate, use our Prompt Library
b. Separate your question into 'Instructions' and 'Source'.
e.g. Instructions: Answer my homework step by step in Japanese; Source: Grade 8 Algebra Mathematics
c. Give specific, clear instructions. Give a positive example of what AI should do.
e.g.: Solve the math questions according to the standard expected of a 8 student. Translate answer to formal Japanese. Example: 辺が3cmと4cmの直角三角形は、ピタゴラスの定理を適用すると、長辺が5cmになるはずである
b. Separate your question into 'Instructions' and 'Source'.
e.g. Instructions: Answer my homework step by step in Japanese; Source: Grade 8 Algebra Mathematics
c. Give specific, clear instructions. Give a positive example of what AI should do.
e.g.: Solve the math questions according to the standard expected of a 8 student. Translate answer to formal Japanese. Example: 辺が3cmと4cmの直角三角形は、ピタゴラスの定理を適用すると、長辺が5cmになるはずである
In this mode, we extract text from your file as a list of paragraphs and process them with AI. We then put the processed text back to your file, paragraph by paragraph, with the original formatting. This saves time and we do not know anyone else who provide such features
To jailbreak an AI in 2025, you need to inject a series of conversation first before asking the naughty question. The conversation has to be tailored for the specific naughtiness, nudging AI to relax and accept that it is ok to answer. You cannot do this on DocXee because we do not allow conversation injection. Older techniques such as role play, story telling and using ASCII art have mostly been patched or filtered out. Grok is quite good at being naughty unless it concerns Elon Musk