There has been limited research on therapeutic alliance in the context of therapist‐assisted Internet‐delivered cognitive behaviour therapy (ICBT) when delivered in clinical practice. The present study investigated therapeutic alliance in ICBT delivered to patients seeking treatment for symptoms of depression (n = 83) or generalized anxiety (n = 112) as part of an open dissemination trial. ICBT was provided by 27 registered therapists or 28 graduate students working in six geographically dispersed clinics; therapist‐assistance was delivered primarily through secure messages and occasionally telephone calls. The Generalized Anxiety Disorder‐7 and Patient Health Questionnaire‐9 were collected pre‐, mid‐ and post‐treatment, and the Therapeutic Alliance Questionnaire was assessed mid‐ and post‐treatment. Therapeutic alliance ratings were high both at mid‐treatment and post‐treatment (above 80%). There was no relationship between therapeutic alliance ratings and improvement on primary outcomes. Among patients treated for depression, lower ratings of mid‐treatment alliance were associated with concurrent treatment by a psychiatrist and fewer phone calls and emails from their therapist. Among patients treated for generalized anxiety, ratings of mid‐treatment alliance were higher among registered providers as compared to graduate students. Multiple directions for future research on therapeutic alliance in ICBT are offered, including suggestions for developing a new measure of therapeutic alliance specific to ICBT and measuring therapeutic alliance throughout the treatment process. Copyright © 2016 John Wiley & Sons, Ltd. Key Practitioner Message This research demonstrated that therapeutic alliance ratings were very strong at both mid‐ and post‐treatment among patients who received Internet‐delivered cognitive behaviour therapy (ICBT) for depression or anxiety in clinical practice. Among patients receiving ICBT for depression, lower ratings of therapeutic alliance were associated with patients reporting concurrent treatment by a psychiatrist and with the receipt of fewer phone calls and emails from the therapist. Among patients receiving ICBT for generalized anxiety, ratings of alliance were higher when patients were treated by registered providers as compared to graduate students. Therapeutic alliance ratings did not predict outcome in ICBT for depression or anxiety. Practitioners have reason to be confident that a therapeutic relationship can be formed in ICBT when delivered in clinical practice.