For CBT and MI trainers, clinical supervisors, counselling programmes
Session fidelity coding, without the tape-coding backlog
Upload a recorded session, apply MITI or your own competence rubric, let the agent take the first pass, then review its calls span by span — and hand the trainee a report that shows exactly which utterance earned which code.
Why fidelity coding is still the bottleneck
Coding a therapy session for fidelity is skilled, slow work. A supervisor listens, tallies behaviours, and writes feedback — for one trainee, one tape. Multiply that by a cohort and the coding becomes the bottleneck that decides how much feedback anyone actually gets.
The usual responses both cost something. Send tapes to an external coding service and you pay per tape and wait. Code them in-house and the work competes with clinical time, which is why so many training programmes rate a sample rather than the whole cohort.
Tagaroo takes the first pass instead. The agent applies the rubric you chose, marks the exact utterance behind every code, and gives its reasoning — so the supervisor's job becomes reviewing and correcting judgments rather than producing them from scratch.
What this costs today
$75
typical cost to have one 20-minute recording MITI-coded with written feedback
20–25%
share of trainee tapes UK CBT courses double-rate to check reliability — the rest go single-rated
NHS Talking Therapies high-intensity CBT curriculum and course handbooks
956
CACREP-accredited US counselling programmes, all assessing recorded practice
Figures checked August 2026. Verify current rates with each source before quoting them.
How it works, end to end
- 01
Bring the recording in
Upload audio or video and Tagaroo transcribes it with speaker attribution, or paste a transcript you already have. De-identify first — a free in-browser de-identifier strips names, dates, and places before anything is uploaded.
- 02
Choose the coding frame
Apply MITI 4.2.1 behaviour counts or the MISC client-language codes from the Scale Library, or upload your own rubric as a PDF and let Tagaroo turn it into an agent skill you can edit.
- 03
Let the agent code first
The agent works through the session utterance by utterance, proposing each code with the span it applies to, how clearly it exemplifies that code, and a written rationale that quotes the criterion it used.
- 04
Review, correct, and disagree
Accept, edit, or reject each proposal. Nothing counts as final until a person signs off, and the record keeps who decided what.
- 05
Measure the agreement
When a second rater codes the same session — another supervisor, or the trainee self-rating — Tagaroo reports agreement, and shows you the specific spans where the two of you diverged. That is the conversation supervision is actually for.
- 06
Hand over the report
Every session produces a printable report: findings with their quotes, severities, and rationales, the agreement summary, and the instrument citation. It is the feedback artefact, not a screenshot of a dashboard.
Instruments included
Each ships with items, anchors, a citation, and its reproduction terms — or bring your own rubric and Tagaroo drafts the agent skills from it.
MITI 4.2.1 — behaviour counts
Nine clinician behaviours: reflections (simple and complex), questions, giving information, persuade, affirm, seeking collaboration, emphasizing autonomy, confront.
View scaleMISC 2.x — client language
Change talk, sustain talk, and follow/neutral, coded against the target behaviour the session is actually about.
View scaleCBT cognitive distortions
For coding the content a client brings rather than the therapist's technique.
View scaleWhat Tagaroo does not do here
- Tagaroo is a research, training, and quality-improvement tool. It is not a diagnostic instrument, and the agent's output is a proposal for a qualified person to review — never an assessment in its own right.
- The CTS-R is not included. It is copyright-protected with permission granted only for individual clinical and supervisory use, so we do not ship an operationalized version; bring your own rubric if your programme rates on CTS-R.
- There is no supervisor role yet: a supervisor runs the study as its admin and trainees join as annotators. Cohort-level dashboards and a designated gold-standard rater are not built.
- Agreement is computed symmetrically between raters. Scoring a trainee against a fixed expert reference standard is on the roadmap, not in the product.
- Identifiable clinical data is not permitted. De-identify recordings and transcripts before upload.
Frequently asked questions
Can Tagaroo code MITI automatically?
It takes the first pass. The agent applies the MITI behaviour counts to each clinician utterance and proposes a code with the exact span, a clarity rating, and a written rationale — then a human reviews every proposal. Published agreement figures for the agent are not available yet, so treat it as a reviewer's assistant rather than a certified coder, and measure it against your own coding on your own tapes.
Is there MITI coding software that does the first pass automatically?
Tagaroo is the closest thing. Upload the session, apply the MITI 4.2.1 behaviour counts from the Scale Library, and the agent proposes a code for each clinician utterance with the span it applies to, a clarity rating, and a written rationale. A human still reviews and signs off every call, and published agreement figures for the agent on MITI specifically are not available yet — so treat it as a first-pass assistant that saves the tallying, not a certified coder.
Does Tagaroo rate CBT competence, not just MI fidelity?
Partly, and the gap is worth knowing before you commit. The Scale Library ships MITI and MISC for motivational interviewing and CBT cognitive-distortion coding for the client's language, but not the CTS-R competence rubric — its licence permits individual clinical and supervisory use only, so we do not ship an operationalised version. If your programme rates on CTS-R, upload your own copy of the rubric and Tagaroo drafts an agent skill for each item, which you then edit.
Do we need a second human rater?
No, but agreement statistics need two complete ratings of the same session. Those can be two supervisors, a supervisor and a trainee self-rating, or a human and the agent — Tagaroo treats the agent as one more rater you can measure rather than as a source of truth.
How much does it cost compared with sending tapes out for coding?
Tagaroo Pro is $29/month, which is less than one 20-minute MITI coding at published external rates, and it covers as many sessions as your credits allow rather than charging per tape. The free plan is for trying the workflow on a transcript you already have — transcribing audio or video, and agreement scoring across raters, are Pro features.
Can we use our own competence rubric instead of a published scale?
Yes, and this is the main reason to pick Tagaroo over a fixed-instrument tool. Upload your rubric as a document and Tagaroo drafts an agent skill for each construct — the operational definition, the examples, the boundary cases, and the severity anchors — which you then edit. The rubric stays a readable file you can put in a supplement or a course handbook.
Is our session data used to train models?
No. Workspace data is hosted in the EU and processed under GDPR, the no-signup trial runs entirely in your browser, and identifiable clinical data is not permitted on the platform at all.
Code one session and see
Start free with a transcript you already have: apply a scale, run the agent, review its calls span by span, and print the report. Your signup credits cover the first passes. Transcribing audio or video is a Pro feature at $29/month — as is agreement scoring across raters.