Brooklyn Trainer
Cambridge-style photo comparison for any topic. Discussion question, optional image prompts, and vocabulary boost — in under 30 seconds.
Open App →Pairs
- 1Pairs sit facing each other with the discussion question projected centrally — this is the Cambridge exam setup and students benefit from replicating the physical arrangement of the real assessment context.
- 2The speaking student should face the images (if used) or the question prompt, while their partner tracks the response against the vocabulary boost words.
- 3Rotate roles after each question so both students in every pair have both the speaking and the monitoring role.
- 4Allow 1 minute for the speaking student to look at images and plan before they begin — exactly as in the Cambridge format.
The Teaching Logic Behind Brooklyn Trainer
Brooklyn Trainer has two modes, chosen by image count. 2 images runs the classic Long Turn (Pair) format: a single speaker compares two photos and answers a discussion question, the Cambridge Speaking Part 2 shape. 3 images switches to the Collaborative Task (Set) format: a small set of images framed around the topic, closer to a Part 3 collaborative discussion than a solo long turn.
Each generation includes a central discussion question on a teacher-specified or custom topic, the AI-generated image prompts for the chosen mode, and an optional Vocabulary Boost — phrasal verbs, key phrases, and idioms at the chosen CEFR level — that the speaker should aim to use naturally during their response.
The comparison aspect is a Cambridge Speaking exam feature: students must not only describe what they see but compare the images, speculate about what is happening, relate the images to the question, and give their opinion. Brooklyn Trainer scaffolds all four of these moves through the vocabulary boost and the comparison aspect field, which specifies the angle of comparison (e.g. 'attitudes to risk', 'impact on the environment', 'how people feel in these situations').
The teacher's role is facilitator and assessor — while one student speaks, the teacher tracks vocabulary boost usage and can use the remaining time to discuss the images with the full class, extending the paired speaking activity into a class discussion when appropriate.
Why It Works
Deliberate practice of exam formats reduces performance anxiety
Dörnyei (2005) identifies L2 anxiety as the most consistently negative predictor of oral exam performance. Regular, low-stakes practice of the exact exam format in a familiar classroom context gradually reduces the novelty of the assessment situation, which is the primary anxiety trigger. Brooklyn Trainer allows weekly Cambridge-format practice without the emotional stakes of a formal test.
Image-based prompts activate speculative language registers
Maley & Duff (1982) demonstrate that images generate a specific type of language that no text prompt can replicate: speculative, interpretive, affective. 'This might be...', 'they appear to be...', 'the contrast between X and Y suggests...' — these language frames only emerge when students must interpret rather than report. This is exactly the register Cambridge photo tasks require.
Pre-loaded target vocabulary changes spoken output
Laufer & Nation (1995) demonstrate that pre-loaded lexical sets — vocabulary students know they are supposed to use — produce significantly more complex spoken output than unprompted free speaking. The vocabulary boost functions as this pre-load: students read the words before they speak, which activates those forms and makes them available for production in the response.
Step-by-Step in Class
Generate the task
Select a CEFR level (B1, B2, or C1), a mode (2 images for the classic Long Turn, 3 for the Collaborative Task), a topic or custom topic, an optional comparison focus, whether to include the Vocabulary Boost, and a speaking time limit (1 to 5 minutes). Review the generated question and vocabulary before displaying to the class.
Set up pairs and assign the long turn
Pair students. One student is the speaker, one is the partner/assessor. Show the images and discussion question. Allow the speaker about a minute to look at the images in silence before starting the clock, exactly as in the Cambridge format — then they have the chosen time limit to compare the photos and address the discussion question.
Partner and teacher assessment
While the speaker talks, the partner tracks vocabulary boost usage (how many of the 3–5 words were used, in what context?). The teacher makes a quick fluency and range note. After the turn, the partner asks one follow-up question — spontaneously, not pre-scripted.
Role switch and class discussion
Switch roles: the former partner becomes the speaker on a new generated question. After both turns, open a class discussion using the comparison aspect as the frame: 'Do you agree with what you heard? What would you have added?' This converts the paired activity into a class speaking task.
How to Set It Up for Different Levels
Generates a B2 First-level discussion question with 2 image prompts and 4 vocabulary boost words including comparison collocations ('in contrast to', 'significantly more likely to'), speculative modals, and evaluative adjectives.
B2 First (Cambridge First) is the most common CEFR-level Cambridge exam in European classrooms. Regular generation of B2-calibrated comparison tasks with image prompts provides direct, low-preparation exam readiness practice across any topic in the curriculum.
Generates a C1 Advanced-level question requiring 3 image comparison with 5 vocabulary boost words targeting hedging ('it is conceivable that'), evaluative nuance ('arguably', 'ostensibly'), and comparative collocations.
C1 Advanced requires candidates to maintain a nuanced extended turn on complex topics — not just compare images but speculate, evaluate, and relate the comparison to broader questions. Three images rather than two requires higher-order integration of multiple viewpoints in a single response.
Generates a B1-level question with 2 simple image prompts but no vocabulary boost — the speaker must manage without pre-loaded vocabulary. Suitable for fluency-focused sessions rather than vocabulary consolidation.
Removing the vocabulary boost creates a different kind of productive challenge: students must use vocabulary they already own without pre-activation support. This measures baseline spontaneous production, reveals actual vocabulary gaps, and provides diagnostic information about individual lexical range.
Ways to Extend the Game
Timed Turn with Vocabulary Tracking
Partner counts how many vocabulary boost words are used in the 60-second turn and reports after. Over a series of sessions, students can track their own improvement in naturally deploying target vocabulary under time pressure.
Follow-Up Question Practice
After the main turn, the partner asks three follow-up questions, not from a list but spontaneously generated from what they heard. This replicates the Cambridge interactive speaking phase and develops the ability to listen actively while formulating questions simultaneously.
Class Consensus Round
After all pairs have completed their turns, the class works together to reach a consensus answer to the discussion question. Each pair contributes one idea from their conversation. The teacher records contributions and facilitates the final shared response.
Image Caption Writing
After the speaking round, students write a 20-word caption for each image using at least one vocabulary boost word per caption. The caption should convey the comparison aspect — this bridges from photo comparison speaking to concise interpretive writing.
Pair It With
Iconic Stories
Iconic Stories provides the narrative dimension of visual interpretation — where Brooklyn Trainer develops comparative speaking from images, Iconic Stories builds story reconstruction and narrative language from a sequence of images.
BubbleTalk
BubbleTalk extends the class discussion phase of Brooklyn Trainer — after paired comparison practice, generate BubbleTalk bubbles on the same topic so one or two students can build a wider spoken discussion at the screen.
Ask, Tell, Reveal
Ask, Tell, Reveal provides the structured spoken follow-up to Brooklyn Trainer's paired long turn — after students have compared images, the ATR game format develops questioning and information-sharing skills at the same CEFR level.