This template library collects the reusable sentence frames, linking words and rule tables from the rest of the course in one place, so you can look something up quickly instead of rereading a whole lesson. Start here: the first section explains how to use every template in the library, and the rest of this lesson covers Describe Image.
How to Use This Library
Find the task you want to work on, learn the template as a sentence scaffold, and then fill every sentence with the real information in the prompt. Never recite the same memorised version word for word. The reason is written into the official scoring rubric: for both Describe Image and Write Email, the lowest Content band says the response "may not deal properly with the prompt due to significant amounts of pre-prepared/memorized material". A response built from memorised material can score 0 for Content, and on those tasks that loses the whole item. The method behind these scaffolds is in Describe Image: structure and content.
This library is a temporary scaffold, not the goal. Early on it lowers the load, because you don't have to think about content and sentence structure at the same time. Once you are comfortable, swap the scaffold sentences for your own wording, and lean on it less as your test gets closer. Two official points explain why:
- Pearson says not to use templated sentences. The official Describe Image page says: "Do not use memorized responses or templated sentences. Such responses can lead to a score of 0 for content."
- Pearson runs a template detector. A 2025 Pearson research paper describes a detection system for templated responses that is already in live use. In testing it produced no false positives and caught a little over half of templated responses. The paper also describes "adaptive adversarial drift": detection rates jump after each model update, then slowly fall as the templates that got caught spread among test takers.
Put together, this means the risk grows with how popular a template is. A widely shared template is exactly the kind that gets caught first. If you can find it online, so can the detector, and that includes the templates below, which many learners will read too.
A suggested way to work through the library:
- First few days (roughly days 1 to 5): practise with the scaffold with one aim only: no freezing. You can start speaking or writing straight away.
- Next stretch (roughly days 6 to 10): ask of each sentence, "Would this still work for a different prompt?" Replace every sentence that would with content that only fits this prompt.
- Last few days before the test: keep only three kinds of stock sentence: linking words, greetings and sign-offs, and opening lines you have already rewritten yourself. Everything else should be your own words by now.
A one-line test: delete every template sentence from your answer. Does what's left still complete the task? If yes, the template was only a scaffold. If no, the template is standing in for your answer, and that is exactly what the rubric scores as 0.
Why Describe Image Needs Real Content
Per Pearson's PTE Core Score Guide, Describe Image is scored on three traits: Content (0 to 5), Pronunciation (0 to 5) and Fluency (0 to 5). Content counts as much as how you sound. The top Content band asks you to describe "all elements of the image and their relationships, possible development and conclusion or implications", while the bottom band is the "pre-prepared/memorized material" wording above.
So the templates below are sentence scaffolds only. The three-part structure and the linking words can stay fixed, but every sentence must carry information that is really on the image: the title, axis labels, actual numbers, the highest and lowest items, and the direction of any trend. You get 25 seconds to prepare and 40 seconds to speak. Check the latest official guide for current timings.
Step One: Identify the Image Type
| Image type | Typical examples | What to focus on | Useful words |
|---|---|---|---|
| Charts (the most common type) | Line graphs, pie charts, bar charts, combined charts, tables | Highest and lowest values, trends | highest, lowest, trend, increase/decrease |
| Maps | Regional distribution maps | Comparing data across areas | area, region, displayed on the map |
| Pictures | Landscape or scene photos | Describing objects and where they are | left/right, central, foreground |
| Flowcharts | Step diagrams, for example seven steps to make coffee | The order of the steps | first step, final step, recycle, repeat |
Decide the type first, then pick the matching template. All four share the same three parts (introduce the image, describe the details, sum up). Only the linking words in the middle change.
Template 1: Charts
Use this for line graphs, pie charts and bar charts.
Opening
The following graph shows information about [title]. The items include [elements, such as the axes or subtitles].
Body
From the graph, we can see that the highest number is in [item], at around [value]. The lowest number is in [item], at just [value]. There is a clear upward/downward trend from [time/item] to [time/item]. When we look at [item], we can see that it is larger/higher/faster than [item].
Every bracket must be filled with something you actually read on this graph: the title, axis labels, real values, the highest and lowest items, the direction of the trend. If you can't fill a sentence with real information, don't say it. Filler that would fit any graph is exactly what the rubric scores as 0 for Content.
Closing
In conclusion, the graph shows that [one sentence on the most striking point in the graph, for example X remained much higher than Y throughout the period].
Templates 2 and 3: Pictures and Maps
Picture opening
The picture shows [subject/scene]. It seems to have been taken in [place/setting].
Picture body
According to the picture, on the left there is [element]. In the centre of the picture, we can see [element]. On the right / At the top / At the bottom of the picture, there is [element], which is big/small/red. [Add one more detail that is really in the picture, such as what people are doing, the weather or the colours.]
Picture closing
In conclusion, the picture mainly shows [one sentence summing up the main subject].
For a map, the body and closing are the same as for a picture. Only the opening changes:
Data for different areas is displayed on the map.
Template 4: Flowcharts
Opening
The following diagram shows information about [title]. It sets out the steps in [process].
Body
From the diagram we can see [elements/subtitle]. According to the diagram, the first step is [element]. The second step is [element]. After that, the third step is [element]. What is more, the next step is [element].
Closing
In conclusion, the process ends with [element].
The Real Red Line Is in the Rubric
The lowest Describe Image Content band in the official Score Guide reads: "Mentions some disjointed elements of the presentation. May not deal properly with the prompt due to significant amounts of pre-prepared/memorized material." If your answer is mostly pre-prepared material, Content drops to 0. Write Email's lowest Content band uses the same wording. Test the scaffolds on Describe Image practice questions.
Note that the Score Guide also says PTE is scored entirely by machine: "human expert scorers are used to train the scoring engine", and once trained, the engine scores every new response. So you don't lose points because a person spots your template. You lose them through the Content rule above. What to do:
- Keep the frame fixed (the three parts and the linking words), but swap every content word for something that is really on this image.
- Test each sentence: would it still be true for a different image? If yes (for example "This image is very useful" or "There are many colours in the picture"), it carries no information. Delete it and say one more real thing about the image.
- Content, Pronunciation and Fluency are each worth 0 to 5. Speak clearly, without stalling, jumping between words or swallowing sounds, and keep your pace even. But if you fill the time with filler instead of what's on the image, Content falls to the bottom band, and good pronunciation won't make up for it.
Respond to a Situation has a similar red line for a different reason. It has no Content trait, but its lowest Appropriacy band describes a response that "contains only a few very short isolated, mainly pre-packaged forms that may not relate coherently to the given situation". Stock phrases that don't fit the situation land in the bottom band there too.