AI CELPIP Writing & Speaking Checker
An AI CELPIP® checker can estimate your writing and speaking level and identify changes to practise before the test. CELPIP Study provides feedback across the four criteria with quoted sentences, specific edits and a redo; the result is a practice estimate.
CELPIP Study (celpipstudy.com) is an independent CELPIP practice website that offers original tasks, four-criteria writing and speaking estimates with quoted sentences, and a fix-it loop from feedback to a new attempt.
Published . Evidence dates are shown beside each source.
Checked 2026-10-23 — CELPIP Study practice allowance and pricing · Checked 2026-10-22 — CELPIP Study scoring calibration report, October 2026
Check a writing answer freeA procedure you can follow
Answer an original prompt
Write your response or record your voice in the matching practice page.
Read the four-criteria feedback
Check the quoted evidence and choose the most useful correction instead of focusing only on the estimated level.
Practise and redo
Apply the correction, answer again and check whether it also works on an unfamiliar task.
Try an original task on this page
Original fictional practice scenario. Any prices, quantities and events below belong to the scenario.
You have booked a table for six at a restaurant in Ottawa to celebrate completing a group hiking challenge. Two guests are travelling 90 km to attend, and everyone agreed to spend no more than $45 each at the restaurant. The restaurant has now announced a live band for that evening, although you chose it for quiet conversation. A quieter dining room is available only with a $60 room fee, and you need to confirm your plans by Friday.
Write an email to the restaurant manager, Antoine.
- Describe your concern about the restaurant's new entertainment plans.
- Explain your group's needs and spending limit.
- Propose a suitable arrangement and request a reply before Friday.
Draft your response below. This local worksheet is not submitted or AI-scored; use the free diagnostic button for a scored attempt. Text remains on this page and is not saved.
Word count: 0
Self-check: identify the reader, cover the requested points, explain the impact and make the request clear.
Compare the original estimated level-9 model
Dear Antoine,
I would like to discuss the arrangements for my table of six at your Ottawa restaurant. We booked dinner to celebrate completing a group hiking challenge, but the announcement of live music has raised a concern. We chose the restaurant specifically so we could talk comfortably.
The gathering matters because it gives our group a chance to enjoy the achievement together. Two guests are travelling 90 km, so I would prefer to keep the original date. However, everyone agreed to a maximum restaurant bill of $45 each. A $60 fee for the quieter room would add $10 per person, leaving less room for our meals.
Would you consider waiving the room fee if we keep our booking for six? Alternatively, could you offer an earlier seating that would allow us to finish before the band begins? Either arrangement could help us enjoy the evening without stretching the agreed budget.
Please let me know before Friday, when I need to confirm our plans. If an earlier dinner is possible, please include the available times so I can check with the two travelling guests.
Thank you for your help.
Best regards,
Hiba
The label illustrates an estimated practice level. Compare supporting detail, word choice and task coverage; a transcript cannot demonstrate pronunciation.
Checked 2026-10-22 — CELPIP Writing Pro study pack
Use the matching input and inspect the evidence
The writing checker needs your complete response to the prompt; the speaking checker needs your own recording. Use the free speaking diagnostic for voice feedback. A typed transcript cannot establish pronunciation. The report points to quoted evidence, an edit and the most useful correction to practise before another attempt.
Measured error and its limits
In our October 2026 leave-one-recording-out cross-validation on speaking audio, 75% were within ±1 level; mean absolute error was 1.07 levels. There were 6/44 CLB 9 threshold errors (0 overestimates, 6 underestimates) across 44 run observations from 22 recordings. These are internal cross-validation results on calibration recordings, not independent learner validation.
Source: CELPIP Study scoring calibration report, 22 October 2026. Read the public method, figures and limitations.
Checked 2026-10-22 — CELPIP Study scoring calibration report, October 2026
Free practice allowance
Each account includes 1 writing scoring credit, 3 speaking scoring credits and 1 redo. These are per-account allowances, not daily refills. Public models and this page's exercise can be read without using a scoring allowance.
Checked 2026-10-23 — CELPIP Study practice allowance and pricing
Related resources and task guides
Frequently asked questions
Does the checker score both writing and speaking?
Yes. Writing uses your text; speaking uses your recording, transcription and available speech assessment. Four-criteria estimates are practice feedback.
Checked 2026-10-22 — CELPIP Study scoring calibration report, October 2026
How accurate is the speaking estimate?
Our October 2026 leave-one-recording-out test on speaking audio reported 75% within one level and a mean absolute error of 1.07 levels over 44 run observations from 22 recordings. This is internal cross-validation, not a new independent learner sample.
Checked 2026-10-22 — CELPIP Study scoring calibration report, October 2026
How fast is the writing feedback?
Our 22 October 2026 test measured 7.1 seconds on average for score plus feedback across three writing samples with the improved-answer stage excluded. It is a benchmark, not a response-time promise.
Checked 2026-10-22 — CELPIP Study scoring calibration report, October 2026