Bring your own key (coming soon)
Claude Sonnet 5.5 for meeting notes on Transcript.audio
How Transcript.audio uses Claude Sonnet 5.5 for meeting notes, where it runs, what it is good at and what to consider. It is used with your own Anthropic account, through bring your own key (coming soon).
What it is used for in the app
Transcript.audio ranks models for each job. Here is where Claude Sonnet 5.5 sits.
Notes and summaries
Rank 6 of 21
Writes the summary, decisions and action items from the transcript.
Structured fields
Rank 5 of 20
Fills template fields such as next steps, owners, budget or dates.
Meeting titles
Rank 6 of 20
Names each meeting so it is easy to find later.
Questions as you go and help
Rank 2 of 20
Answers questions about the meeting while it is happening, and in-app help.
- Best quality
- Recommended
Strengths and trade-offs
- Notes and summaries: Nearly as good as Opus at half the price, and a little quicker.
- Structured fields: Nearly as good as Opus at half the price, and a little quicker.
- Meeting titles: Nearly as good as Opus at half the price, and a little quicker.
- Questions as you go and help: Nearly as good as Opus at half the price, and a little quicker.
Privacy: where it runs
It is used with your own Anthropic account, through bring your own key (coming soon). Compare privacy optimised and standard mode.
- How it connects
- With bring your own key (coming soon) you connect your own Anthropic account, so your usage is billed by Anthropic directly and runs under your own agreement with them.
- Recordings
- Saved on your computer.
- Transcripts and notes
- Saved on your computer. Nothing is stored by us.
- Internet needed
- Yes, for transcription and notes.
Prefer nothing to set up? Transcript.audio cloud (standard) is included in your plan: The transcript is sent to our AI service, summarised by an AI provider under zero data retention, then discarded.
Speed, accuracy and cost
- Notes and summariesAccuracy 4/5Speed 3/5Cost: About $0.05 an hour
- Structured fieldsAccuracy 4/5Speed 3/5Cost: About $0.04 an hour
- Meeting titlesAccuracy 4/5Speed 3/5Cost: About $0.03 an hour
- Questions as you go and helpAccuracy 4/5Speed 3/5Cost: About $0.04 an hour
Accuracy and speed are our own 1 to 5 ratings from the model catalogue, not benchmarks. Costs are approximate provider costs in US dollars per hour of meeting. With your own key, the provider bills you directly.
Presets that pick it
Recommended
The best balance of accuracy and cost.
- Notes and summaries: fallback 4
- Structured fields: fallback 3
- Questions as you go and help: fallback 1
Most accurate
The top model for each task, whatever it costs.
- Structured fields: fallback 3
- Meeting titles: first choice
- Questions as you go and help: fallback 1
Presets pick a model for every job in one click, and fall back down their list when a model is not available. See all presets.
Plan availability
Claude Sonnet 5.5 is used with your own API key. Bring your own key is coming soon as a free update. About bring your own key.
Frequently asked questions
What does Claude Sonnet 5.5 do in Transcript.audio?
Transcript.audio ranks Claude Sonnet 5.5 for notes and summaries (ranked 6 of 21); structured fields (ranked 5 of 20); meeting titles (ranked 6 of 20); questions as you go and help (ranked 2 of 20). Nearly as good as Opus at half the price, and a little quicker.
What happens to my meeting data with Claude Sonnet 5.5?
How it connects: With bring your own key (coming soon) you connect your own Anthropic account, so your usage is billed by Anthropic directly and runs under your own agreement with them. Recordings: Saved on your computer. Transcripts and notes: Saved on your computer. Nothing is stored by us. Internet needed: Yes, for transcription and notes.
How much does Claude Sonnet 5.5 cost?
The approximate provider cost is about $0.05 an hour of meeting, billed to you directly by Anthropic when you use your own key.
Which presets use Claude Sonnet 5.5?
Recommended and Most accurate. Presets pick a model for every job in one click and fall back down their list when a model is not available.
What are the alternatives to Claude Sonnet 5.5?
The models ranked next to it are Gemini 3.1 Pro, AWS Bedrock (Automatic), GPT-5.6 Sol and gpt-oss-120b on Groq. Each has its own page with the trade-offs.