Your creator profile
Your creator profile is byline's working description of your voice, themes, audience, and formats. Supported connected tools can read it as context for your work; the tool decides when to request and use that context.
What's in your profile
byline distills your profile into a handful of facets, each a plain description of one part of how you work:
- Topics: the subjects you actually cover.
- Voice: how you sound. Tone, word choice, sentence shape, the phrases that are unmistakably yours.
- Recurring themes: the throughlines and beliefs that keep surfacing across your work.
- Audience: who you are talking to, and what they keep asking for, drawn from your content and any comments imported earlier.
- Formats: the structures that work for you, from teardown reels to one-idea newsletters.
- Cadence: your posting rhythm, and how an idea tends to develop over time.
Built from your own work
byline builds the profile from a sample of your own processed work, including private material you choose to add. It is a model-generated interpretation and can be incomplete or wrong. Other people's public work is not part of your writing voice.
You can add corrections and pinned facts, such as "based in Lisbon" or "never uses emojis". These are authoritative inputs for your context, but you should still check how a connected tool applies them.
Check the interpretation against your work
Compare a profile claim with the original posts or transcripts in your Library. Not every facet has a direct, per-claim citation in the dashboard. If a description does not fit, correct it rather than treating the profile as proof.
It refreshes overnight
Scheduled consolidation uses available processed My content to refresh the profile. It can take a later run for a newly connected backlog to contribute. Payment, reading limits, processing failures, and insufficient material can delay a refresh; connecting a source alone does not mean the profile is ready.
Write in your voice
A supported AI host can request your byline profile, but it will not necessarily use it in every draft. To connect, use that host's current custom remote MCP option (if available on your host plan), enter the byline endpoint, and complete its authorization flow:
https://byline.fyi/api/mcp
Select the intended byline workspace, then explicitly ask the connected tool to retrieve your profile. Reconnect and authorize again if a saved grant no longer works. Host menus vary; see the connection guide.
Ask in plain language, for example "read this client's byline profile, then draft this announcement in their voice." Check the selected workspace and review the draft against the client's work.
See the difference
In your own connected tool, try the same prompt with and without the client's byline context. Compare the results yourself and ask for supporting sources where available. This is a comparison you run in your tool, not a promised personalized comparison generated by the dashboard.
How long until it sounds like you
One substantial channel can provide a useful starting point, but there is no guaranteed time or amount of content that makes a profile accurate. Check processing status in the Library, inspect the available profile, and test a specific recall question before relying on it for a client draft.
Reading-status filters work while browsing the Library. Clear search before choosing a reading status. Inspecting stored Library text requires current paid or complimentary access and shows up to 20,000 characters; it is not a full-data export.