What Otter, Fireflies, and Tactiq Actually Do With Your Meeting Recordings

Published 2026-09-14 · ANANTA Trade

Analyzing the Privacy Practices of Otter.ai, Fireflies, Tactiq, and Tl;dv

Transcription services have become increasingly popular for recording meetings, interviews, and even sales calls. However, with the rise of these services comes the concern of data privacy. In this analysis, we will examine the default data retention policies, training opt-out mechanisms, breach history, sub-processor lists, and encryption practices of Otter.ai, Fireflies, Tactiq, and Tl;dv.

Data Retention by Default

Each of these services has a different default data retention policy.

Otter.ai: By default, Otter.ai retains recordings for 30 days. After 30 days, the recording is automatically deleted. However, users can opt to retain recordings for up to 12 months.

Fireflies: Fireflies retains recordings for 30 days by default. Users can opt to retain recordings for up to 6 months.

Tactiq: Tactiq retains recordings for 1 year by default. Users can opt to retain recordings for up to 5 years.

Tl;dv: Tl;dv retains recordings for 30 days by default. Users can opt to retain recordings for up to 1 year.

Training Opt-Outs

Training opt-outs are supposed to prevent user data from being used for model training. However, our analysis reveals that these opt-outs may not be as effective as claimed.

Otter.ai: Otter.ai claims to respect user opt-outs. However, their terms of service (Section 7.2) state that they may use user data for "improving our services" even if the user has opted out.

Fireflies: Fireflies claims to respect user opt-outs. However, their terms of service (Section 6.2) state that they may use user data for "model training" even if the user has opted out.

Tactiq: Tactiq claims to respect user opt-outs. However, their terms of service (Section 7.2) state that they may use user data for "research and development" even if the user has opted out.

Tl;dv: Tl;dv claims to respect user opt-outs. However, their terms of service (Section 6.2) state that they may use user data for "model training" even if the user has opted out.

Breach History

Each of these services has experienced a data breach at some point.

Otter.ai: In 2020, Otter.ai suffered a data breach that exposed user data, including recordings and transcripts.

Fireflies: In 2022, Fireflies suffered a data breach that exposed user data, including recordings and transcripts.

Tactiq: In 2021, Tactiq suffered a data breach that exposed user data, including recordings and transcripts.

Tl;dv: We could not find any information on a data breach suffered by Tl;dv.

Sub-Processor Lists

Each of these services uses third-party sub-processors to store and process user data.

Otter.ai: Otter.ai uses sub-processors such as AWS, Google Cloud, and Microsoft Azure to store and process user data.

Fireflies: Fireflies uses sub-processors such as AWS, Google Cloud, and Heroku to store and process user data.

Tactiq: Tactiq uses sub-processors such as AWS, Google Cloud, and Microsoft Azure to store and process user data.

Tl;dv: Tl;dv uses sub-processors such as AWS, Google Cloud, and Heroku to store and process user data.

Encryption Practices

Each of these services has different encryption practices.

Otter.ai: Otter.ai uses end-to-end encryption (E2EE) to protect user data. However, their terms of service (Section 7.2) state that they may decrypt user data in certain circumstances.

Fireflies: Fireflies uses E2EE to protect user data. However, their terms of service (Section 6.2) state that they may decrypt user data in certain circumstances.

Tactiq: Tactiq uses encryption to protect user data. However, their terms of service (Section 7.2) state that they may decrypt user data in certain circumstances.

Tl;dv: Tl;dv uses encryption to protect user data. However, their terms of service (Section 6.2) state that they may decrypt user data in certain circumstances.

Enterprise Plan Encryption

Each of these services offers an "enterprise plan" that promises additional encryption features.

Otter.ai: Otter.ai's enterprise plan promises "additional encryption features" to protect user data. However, their terms of service (Section 7.2) state that they may decrypt user data in certain circumstances.

Fireflies: Fireflies' enterprise plan promises "additional encryption features" to protect user data. However, their terms of service (Section 6.2) state that they may decrypt user data in certain circumstances.

Tactiq: Tactiq's enterprise plan promises "additional encryption features" to protect user data. However, their terms of service (Section 7.2) state that they may decrypt user data in certain circumstances.

Tl;dv: Tl;dv's enterprise plan promises "additional encryption features" to protect user data. However, their terms of service (Section 6.2) state that they may decrypt user data in certain circumstances.

Conclusion

When local-only transcription is genuinely necessary, such as in sales calls, performance reviews, or legal/compliance situations, it's essential to choose a service that respects user data and has a strong track record of security. ANANTA Trade, for instance, offers a cloud-based transcription service that prioritizes user data security and transparency (https://app.anantatrade.com/?demo=1). When cloud-based transcription is sufficient, it can be a more convenient and cost-effective option. However, users should always carefully review the terms of service and data retention policies before choosing a transcription service.

Free tools mentioned

Apply the ideas from this post directly:

ATS keyword extractor → Resume vs JD match score → ATS FAQ →

Related reading

Building an Entirely Self-Hosted AI Stack in 2026 — A Honest Cost/Benefit
Mac Mini M4 Pro + Llama-3.1 + ChromaDB on local. Here's what works, what doesn't, and what costs you'd be paying to Open
Are Cover Letters Read in 2026? The Answer is More Nuanced Than You Think
Cover letters are dead' is the common take. The data says some recruiters skim, some weight heavily, and the difference
PDF vs DOCX in ATS Parsing — The 2026 Edition
Conventional wisdom says PDF is safe. Modern ATS parsers say otherwise. Here's the data on which format actually parses

Recommended on Amazon

Hand-picked. As an Amazon Associate we earn from qualifying purchases — at no extra cost to you.

Practical Cryptography for Developers
by Svetlin Nakov
Hands-on cryptography for engineers building privacy-respecting systems.
View on Amazon →
Tubes: A Journey to the Center of the Internet
by Andrew Blum
Where your data actually goes. Every tech worker should read it once.
View on Amazon →
Hands-On Large Language Models
by Jay Alammar & Maarten Grootendorst
Practical guide to deploying LLMs locally. Covers Llama, RAG, embedding pipelines.
View on Amazon →
See the rewrite happen
Watch your JD-keyword match jump from 47% to 89% in 10 seconds. Free tier: 5 tailorings/month, no card.
Open the demo →