# How to Use Interpreter with Agency Portals and Softphones > Agency portals and softphones already run your OPI day. Learn how to add Interpreter without disrupting call controls, audio, or privacy. Author: Mirkamol Published: May 5, 2026 Updated: September 8, 2026 Categories: Guides Canonical page: https://useinterpreter.com/blog/interpreter-agency-portals-softphones/ Most OPI interpreters do not control the main platform. The agency tells you which portal to use. The client call arrives through a softphone, browser queue, remote desktop, or meeting link. You sign in, set your status, accept the call, and work inside whatever system the assignment uses. A support tool has to fit around that reality. If it asks you to replace the agency portal, invite the caller somewhere else, or change the call flow, it will not survive a busy shift. [Interpreter](/) is designed to sit beside your portal or softphone. You keep the call where it already is. Interpreter listens to the approved audio source and gives you live transcription, two-way translation, notes, and lookup on screen. For role and conduct basics, see the [NCIHC standards of practice](https://www.ncihc.org/standards-of-practice). The portal is where the call starts. Keep its controls within reach when adding anything beside it. ## Start with policy Before setup, confirm that your agency or client allows live transcription support for your call type. > [!TIP] > Do not cover the softphone controls with notes or captions. Mute, hold, and disconnect need to stay visible. Ask whether the tool is approved for: - General customer service calls - Medical calls - Legal calls - Government or benefits calls - Insurance and finance calls Some agencies allow support tools for general calls but restrict them for healthcare or legal work. Some require a written approval. Some do not allow outside tools on agency-owned computers. Use the policy you receive. If you need a cautious overview, read [Is real-time transcription allowed for interpreters?](/blog/real-time-transcription-allowed-for-interpreters/). ## Keep the agency portal as the call source Do not move the call into Interpreter. Keep your normal portal or softphone in charge of: - Accepting calls - Muting and unmuting - Dialing out - Call transfers - Hold - Disposition codes - Time tracking Interpreter should support the live work, not replace the operational system. Place it next to the portal on your screen so you can see the transcript without hiding call controls. If you use one monitor, keep the portal and Interpreter in split view. If you use two monitors, put the portal and transcript where your eyes can move with little effort. Avoid placing notes on a far edge of the screen where you will not use them. ## Browser portals For a browser portal approved for Chrome or Edge, start Interpreter in another tab or window. In the sharing dialog, select the source that actually offers the call audio. Available audio-sharing options vary by browser, operating system, and device policy. Make sure the browser shares audio. This step matters. A screen share without audio gives Interpreter no call sound to transcribe. Then test with safe audio. You should see text appear when the portal audio plays. If you see only your own voice, check whether you selected the microphone instead of the call audio. For more detail, read [How to share call audio for live interpreter captions](/blog/share-call-audio-live-interpreter-captions/). ## Desktop softphones Desktop softphones may need system-audio capture, where your browser and operating system support it. Keep your softphone output on your headset and its input on your microphone. Test the approved capture path before accepting calls. Do not change softphone settings mid-shift unless you have time to test. A small device change can break inbound calls or create echo. If your softphone offers a browser version, ask whether your agency allows it. Browser-based audio can make capture easier, but policy decides. [Audio Routing for OPI Interpreters](/blog/audio-routing-opi-interpreters/) walks through the common setups. ## Remote desktops and locked-down computers Some agencies route calls through a remote desktop or a locked-down work machine. In those environments, you may not be able to share system audio with a local browser tool. Do not work around that restriction with personal devices unless your agency approves. Sensitive call audio should not pass through a phone speaker, personal recorder, or unapproved app. Ask the agency whether it supports a compliant caption or transcription workflow. If the answer is no, use approved note-taking habits and clarification scripts instead. You can still use Interpreter for call types and devices where approval exists. ## Set up the Interpreter workspace Before calls start, review the available domain and glossary settings. Use general terminology relevant to your assignments, not patient names or identifiable case details. A vocabulary hint may help recognition; it does not guarantee the right word. Keep floating notes narrow. Use them for temporary details like appointment times, case numbers, callback numbers, and addresses. Do not turn them into a retained record of the call. Use quick lookup for terms you need during the session. Keep it inside the workspace so you are not jumping between a portal, search engine, and notes app while someone is speaking. ## During the call Let the portal run the call. Let Interpreter support accuracy. Use the transcript as a second reference, then confirm critical details against the audio or the speaker. Speaker labels can also be wrong. Use approved notes for fragile numbers and lookup for general terms without patient context. If audio gets unclear, ask for clarification through the parties. Do not let the transcript guess replace what you hear. The interpreter remains responsible for the rendering. ## After the call End the call in the agency portal first if that is your normal workflow. Complete any disposition steps. Then close the Interpreter session. Do not copy sensitive transcript text into personal notes. Do not save screenshots unless an approved process requires it. If you used temporary notes, clear them according to policy. Have the agency review [Interpreter's privacy information](/privacy/), actual data flows, and applicable agreements before approval. Do not infer that ending a session deletes every copy or that a no-recording policy permits third-party audio processing. ## Make the setup repeatable The best portal-plus-Interpreter workflow is routine. Same browser. Same headset. Same screen layout. Same audio source. Same pre-shift test. Once the setup becomes predictable, the tool fades into the background. The portal keeps handling calls. Interpreter keeps the words visible. You keep interpreting. --- *Related reading:* - [Teleperformance interpreter tools: an independent setup guide](/blog/teleperformance-interpreter-tools/) - [Audio routing for OPI interpreters](/blog/audio-routing-opi-interpreters/) - [No-recording interpreter notes and privacy](/blog/no-recording-interpreter-notes-privacy/) - [The interpreter's toolkit](/blog/interpreter-tools-guide/) - [Interpreter pricing](/pricing/)