What prompt makes a history AI useful?
A history AI gives a better answer when you tell it where to look. Paste this into Claude, ChatGPT, Gemini or Whizi and fill the brackets. It asks for a sourced answer, and it marks where historians disagree.

Explain [event or question] in [country] between [start year] and [end year]. Search the web and use only [source types, such as museum and archive sites or university presses]. Give a three sentence answer first. Then list each date, name and number with a link beside it. Mark any claim where historians disagree, and name the two main views.
Each bracket shuts a door. The place and dates stop the model from mixing two events with the same name. The source types keep a content farm from outranking an archive. The last line is the one that matters for essays, because a history question usually has more than one defensible answer.
Which history AI is free?
Most history tools let you start free, and the three Google shows first for this query work in different ways. The table lists what each says on its own page, read on 22 November 2026. Where a page said nothing, the cell says "not stated" instead of a guess.
| Tool | Free use | Signup | Paid price | Sources shown | Best at |
|---|---|---|---|---|---|
| Edubrain | Homepage FAQ says free, no attempt limits | Homepage FAQ says no login required | $3.99 a week for AI-Plus; expert help from $3.99 | Not stated for the history page | Fast homework answers with steps |
| HelloHistory | Page shows two offers, 20 and 30 free messages, above the app store buttons | Not stated; links go to app stores | None listed | Not stated | Talking to a historical figure in an app |
| World History Encyclopedia | All content free | Not stated | Member subscriptions exist; price not stated | Peer reviewed articles by named scholars | Reading a vetted article |
| ChatGPT | Free $0; search listed; uploads "limited" | Not stated | Go $8, Plus $20 a month | Search results with links | One chat for questions, files and drafts |
| Claude | Free; limited usage every 5 hours and every week | Not stated | Pro $20 a month, or $17 a month billed annually | Search listed on Free | Long, careful essays and primary source readings |
| Whizi | 10 messages a day on Whizi AI, Auto and Claude Haiku, no card | Account needed | Pro $29.99 a month | Web search on by default | Claude, GPT and Gemini in one app |
Whizi loses on three counts. Edubrain says you need no login, and Whizi needs an account. World History Encyclopedia is a vetted reference that a chatbot can't replace, so read it and check any date against it. And HelloHistory is built for talking to a historical figure, where Whizi gives you a general chat you set up with a role prompt.
The page lists no accuracy scores for any of them. Edubrain's history helper and HelloHistory's app page publish no accuracy figures that I could find, so treat their answers as drafts. For a wider look at study tools, see study AI.
How do you use AI for history? Three worked examples
Each example below is written for this page to show the shape of a good prompt and a good answer. None is a recorded output from a named model.
- A primary source. Prompt: "Here is a passage from a 1919 newspaper editorial on the Treaty of Versailles: [paste]. Who wrote it, for whom, and why? Separate what the author states from what the author wants the reader to feel. Don't tell me what the treaty did unless the passage mentions it." A good answer names the author's position, points to loaded words in the passage, and says what the source can't show, such as how most readers felt.
- A causes essay. Prompt: "My thesis is that the Roman Republic fell mainly because armies became loyal to generals, not the state. Give me the three strongest objections a historian would raise, each with one named scholar or primary source I can look up. Mark any you aren't sure exists." A good answer pushes back with real arguments, gives names you can search, and flags any it can't source. If a name has no results, drop it.
- A date check. Prompt: "I wrote that the Berlin Wall fell in 1991. Check it against a reliable source, give the link and the exact date, and tell me what happened in 1991 so I don't confuse the two." A good answer corrects the year, points to a page you can open, and names the other event. Your job is to open the page and find the date.
In all three, the model works from a source, shows where it got each claim, and leaves the judgment to you.
What are 12 prompts for history homework and essays?
These twelve are grouped by the job. Replace each [bracket] with your own detail and they work on the first try.
Understand an event:
- Timeline: "Make a timeline of [event] from [start] to [end] with a link for each entry. Use only [source types]. Mark any date that sources give differently."
- Causes and effects: "List the short term and long term causes of [event], then the effects. Say which cause historians weigh most and why, and name one who disagrees."
- Two sides: "Give the strongest case for each side in [debate, such as whether the New Deal ended the Depression], three sentences each, then say what evidence would change the answer."
Read a source:
- Source analysis: "Here is a source: [paste]. Tell me author, audience, purpose and bias in four lines. Then list two things it can't tell us."
- Compare accounts: "Here are two accounts of [event]: [paste both]. Where do they disagree on facts, and where only on tone? Which would you trust for the date and why?"
- Summarize a reading: "Summarize this chapter in ten bullets: [attach PDF]. Use only the chapter, and put the page number beside each bullet."
Write and revise:
- Thesis test: "Here is my thesis: [paste]. Is it arguable, or just a fact? Rewrite it twice, once narrower and once with a clearer reason, and keep my meaning."
- Outline: "Build a five paragraph outline for this question: [question]. Each paragraph gets a claim, two pieces of evidence I should find, and a line linking it back to the question. Don't write the paragraphs."
- Feedback: "Read my paragraph: [paste]. Where is a claim with no evidence? Where am I describing instead of arguing? Point to the line and stop. Don't rewrite it."
Revise and practice:
- Exam questions: "Write six essay questions on [unit] at [course level], two on cause, two on change over time, two on comparison. Give a marking guide of four points for each."
- Quiz on dates: "Quiz me on these dates one at a time: [paste list]. Ask for the event first, then the year. Repeat the ones I miss at the end."
- Talk to a figure: "Act as [historical figure] in [year], speaking only about what a person in that place could know then. Answer my questions in character. After each answer, add a line in brackets saying which part is documented and which you are inventing."
Prompt 12 covers the "talk to historical figures" apps. The bracket line is the safety valve: a character answer reads like fact unless the model labels the guesses.
How do you read a primary source with a history AI?
Paste the document yourself and ask questions about that text only. A model that has the passage in front of it has less room to fill gaps from memory, and you can check every answer against the lines you pasted.
Here is a primary source: [paste the text and say where it came from]. First, tell me who wrote it, when, and for what audience, using only what the passage says. Second, quote the two phrases that show the author's purpose. Third, list what the passage can't tell us and what other source would fill that gap.
Three habits make this work. Paste the whole passage, because a clipped quote hides the sentence that changes its meaning. Ask for exact phrases, then find them in your text: a phrase that isn't there means the model made it up. And keep the context in your own hands. The model can suggest that a source was written to persuade, but who held the pen and why is something you confirm in the source's own introduction or a reference work.
Then ask for a second reading from the other side: "Who would have disagreed with this author in the same year, and what would they have said?" Treat the answer as a lead to look up, not as a fact, and cite the original document in your essay.
Can a history AI do my homework?
It can do the parts that build understanding, and it shouldn't do the part your teacher grades. A map question, a vocabulary list and a practice quiz are fair uses. A finished essay handed in as yours is not, and most schools say so in their academic honesty policy.
The safer split is simple. Let the model ask you questions, check your timeline and find the weak spot in your thesis. Then close the chat and write. If you get stuck on a source, ask the model to explain what a word meant in that period, not to write your paragraph. A teacher can tell when a paragraph knows more than you do.
For other subjects the same split works. AI homework helper covers photo and problem sets, and science AI covers lab and paper questions.
What mistakes ruin a history AI answer?
Four mistakes cause most of the damage, and each has a quick fix.
- Trusting a date with no link. The fix: ask for a source beside every date and open it. If the page doesn't say it, drop it.
- Asking for names of scholars and books, then citing them blind. The fix: search each name before it goes in your bibliography. Invented citations look real, as the guide to fact checking AI answers shows.
- Treating a role play as a source. The fix: use a character prompt for practice only, and cite the documents it drew on.
- Pasting the answer into graded work. The fix: use the model to test your argument, write in your own words, and cite the original source.
A fifth is the quiet one: no period. "The revolution" could be four countries, and the model picks the likeliest one without telling you.
Is AI accurate for history?
Check every exact date, quote and attribution before you use it, because a model predicts likely text and a likely sentence isn't always a true one.
Google gave a public example. In a post dated 23 February 2024, Google said the tuning that made Gemini show a range of people "failed to account for cases that should clearly not show a range", and that it did not want the model to create inaccurate historical images. It said it had temporarily paused image generation of people. The post concerned pictures, and the same lesson applies to text: the answer looks plausible, and only the source tells you.
Check your school's rules before you use any of this on graded work, because many set limits on AI help. For a deeper look at why models state wrong facts, see why AI gets things wrong.
Where does Whizi fit as a history AI?
Whizi is one app for history questions from Claude, GPT and Gemini, instead of a pile of single purpose tools. Web search is on by default in a new chat, so dates can come with sources. You can attach a PDF reading, a Word file or a scan, and scanned PDFs get a vision transcription pass first.
The free plan gives you 10 messages a day on Whizi AI, Auto and Claude Haiku, with no card. Pro carries Claude, GPT and Gemini in one app, and its first week is $1.99. Side by side Compare, which puts two models on one prompt, is a Pro and Powerhouse feature, and it is handy for prompt 2 or the date check, because you read only where the two answers disagree.
Pin your syllabus and chapters to a Project so each chat starts from them. A Word file made in chat, such as an essay outline, has no plan gate. Edubrain and World History Encyclopedia are still quicker for a one off question or a vetted article.
What can a history AI not do?
It can't replace the archive, the textbook or your own reading of a source. It can't tell you what your teacher expects, and it can't vouch for a date it can't link to.
It also tends to smooth over disagreement. Real history is full of questions where scholars differ, and a chat answer often picks one side in a confident voice. Prompts 2 and 3 make it show both. When the answer matters for a grade, a degree or a published piece, check it against a vetted reference such as a university press book or an encyclopedia whose articles are reviewed.
- Name the place and the period in every history prompt
- Switch web search on and ask for a link beside every date
- Search each scholar or book the model names before you cite it
- Label any role play answer as practice, not a source
- Write the argument yourself and cite the original source
Frequently asked questions
Is there a free history AI with no login?
Yes. Edubrain says on its site FAQ that you can use it for free with no attempt limits and no login required, as of November 2026. The page gives no accuracy data for its history tool, so check any date it gives you against a reference.
Can AI talk to historical figures?
Yes, with limits. HelloHistory describes an app for conversations with historical figures, and a chatbot can do it from a role prompt like prompt 12. Treat every reply as fiction built on sources, and ask the model to mark which parts are documented.
Can AI help with a history essay?
It can test a thesis, find objections and give feedback on a draft. It shouldn't write the essay, and a school may forbid it. Use prompts 7 to 9, then write in your own words and cite the original sources.
What is the best AI for history?
No test ranks them for history, so pick by the job. Claude and ChatGPT both list web search on their free plans, as of November 2026, and World History Encyclopedia is the one source in the table whose articles are peer reviewed. Use a chatbot for questions and drafts, and the encyclopedia for facts.
Can AI read a history textbook PDF?
Usually yes. ChatGPT lists "limited messages with uploads" on its free plan, as of November 2026. Attach the chapter, tell the model to answer only from it and put a page number beside each point, as in prompt 6.