The spread of electronic health records systems (EHRs) poses challenges for both patient and provider care experience. Limited research suggests that scribes offer potential benefits to productivity and clinician satisfaction in emergency health and specialty settings. We conducted this evaluation of trained volunteer scribes for primary care clinics serving a diverse, low-income population in a US safety net system, which implemented a new EHR 2011–2014. The scribe programme trained and managed scribes for 51 providers (25% participation) from 5 of 12 San Francisco Health Network primary care clinics. We evaluated the programme using four measures. Providers reported spending less time out of clinic completing notes after sessions with scribes versus sessions without scribes (14.0 min vs 30.2 min, p<0.01). The rate of incomplete EHR notes at 72 hours was not significantly different for clinics using and not using scribes (16.9% vs 16.7%, p=0.4). Mean visit length using EHR-recorded provider cycle time was shorter for sessions with scribes (24.0 min), compared with sessions without scribes (26.4 min, p<0.01). Patients at clinics using scribes were as likely to recommend their provider (74.5%), compared with patients at clinics not using scribes (74.3%). Limitations of our evaluation include selection bias and possible confounding by clinic- and provider-level factors. In a safety net primary care system, trained volunteer scribes were associated with improved clinician efficiency and experience and no difference in patient satisfaction.