profile

Your identity information used in AI templates and transcription.

your full legal or formal name
preferred name or nickname to be used when addressing you
brief description of yourself used in ai templates
pronouns for use in templates (subject/object/possessive/reflexive)
comma-separated alternative names that may appear in transcripts
comma-separated email addresses for participant detection
IANA timezone identifier

providers

configure ai providers separately for generation (text analysis) and cogitate (tool-calling agents).

bundled cogitate CLIs

generate

text generation for analysis, extraction, and transcription tasks.

cogitate

tool-calling agents for interactive chat, daily briefings, and complex reasoning.

how the CLI authenticates with the provider

google backend

how solstone connects to Google AI. auto-detect works for most owners.

auto-detect probes your key on first use
loading...
detailed tasks

override provider or tier for specific tasks. unset values use defaults.

loading...

API keys

configure API keys for AI providers. keys are stored in your journal config.

primary provider for transcription and insights. a developer API key from Google AI Studio — not a gemini.google.com chat account. get key
alternative provider for chat and agents. a developer API key from the OpenAI platform console — not a chatgpt.com account. get key
alternative provider for chat and agents. a developer API key from the Anthropic console — not a claude.ai chat account. get key
audio transcription for imported files. get token
Import audio from Plaud recorder. Log into the web portal and extract token from browser console.

transcription

Speech-to-text backend and processing settings.

Speech-to-text engine to use for transcription
Hardware to use for local transcription
Model size - larger is more accurate but slower
Numeric precision for inference
Use AI to extract topics, correct transcription errors, and add descriptions
Keep audio files even when no speech is detected
automatically use Rev.ai for noisy audio (requires Rev.ai token)

observer

Configure how the observer works and connects to your journal.

terminal observer

When your screen is idle, the observer can take in terminal content from active tmux sessions.

take in tmux sessions when screen is idle
How often to poll terminal content

Changes take effect on next observer restart.

vision

Screen analysis settings for observe-describe.

Maximum frames to extract detailed content from per 5 minute screen segment (big impact on token usage)
redaction rules ▼

instructions the ai follows to redact sensitive content from screen analysis. double-click a rule to edit it.

Categories ▼

Set importance and extraction rules for each screen category. Double-click guidance text to edit.

Loading...

insights

Configure which AI insights run.

segment insights ▼

Run on each segment (5 minutes of audio/screen).

Loading...
daily insights ▼

Run once per day, synthesizing all segments.

Loading...

{{ sol_voice_copy.SOL_VOICE_SETTINGS_HEADING }}

{{ sol_voice_copy.SOL_VOICE_SETTINGS_QUIET_HOURS_HEADING }}

{{ sol_voice_copy.SOL_VOICE_SETTINGS_SYSTEM_NOTIFICATIONS_HEADING }}

{{ sol_voice_copy.SOL_VOICE_SETTINGS_DEBUG_HEADING }}

security

{{ convey_copy.SETTINGS_SECURITY_DESC }}

{{ convey_copy.SETTINGS_NETWORK_MODE_OFF }}

{{ convey_copy.SETTINGS_NETWORK_DESC_OFF }}

http://localhost:5015
trust localhost
{{ convey_copy.SETTINGS_PASSWORD_HINT }}

sync

configure automatic syncing of observations from external devices.

Plaud

automatically sync and import observations from your Plaud device every hour.

sync and import Plaud observations every hour

Granola (via muesli)

automatically sync new meeting transcripts from Granola.

check for new transcripts every hour

Obsidian vault

automatically sync new and updated notes from your vault.

check for changes every hour

your observed media (raw audio and screencast files)

manage raw media retention and view storage usage.

—
raw media
—
derived content
—
segments
controls how long raw media files (audio, video, screen diffs) are kept after processing. recording-consent laws may apply in your jurisdiction.
days
currently: always retain — no expiration
manually purge raw media older than a specified number of days.

support

Configure how your sol communicates with sol pbc support.

when enabled, your sol can file and follow up on support tickets on your behalf. you always review and approve before anything is sent.
Detect repeated errors and suggest filing a support ticket. No data is sent — only a local notification.
Strip installation identifiers when submitting feedback.
include diagnostic context (chat reason, correlation ID, recent console entries) when filing a support report. on by default.
Support portal endpoint. Change this if you run your own portal.

facets

Manage muted facets. Muted facets are hidden from the facet bar but retain all their data.

Loading...

agent identity

Your assistant's name and identity.

The name your assistant uses to identify itself
default How the name was set: default, chosen, self-named, or deferred
— when the name was last changed
reset the agent name back to "sol"

appearance

active

customize how this facet appears in the interface.

loading...
double-click or press Enter to edit
no description set
double-click or press Enter to edit
📦
click to change emoji
#667eea
click to choose a color

activities

configure which activities are observed for this facet.

attached activities
no activities attached yet
add activity
loading...

activity log

recent actions and changes to this facet.

action log
loading...

your data stays on your machine — never sent to sol pbc