Settings
Seven tabs, one page. Everything from AI model selection to team invites to emergency stops. Settings are per-company — each company you belong to has its own configuration, billing, and member list.
General
Company name, custom instructions, and the controls that let you pause or stop the AI.
Custom Instructions
Persistent rules the AI follows across all sessions. Set them here or tell Solmyr in chat — either way they stick. Ideal for communication preferences, workflow constraints, or standing orders like "check in every 30 minutes."
Company name
Rename your company anytime. This is the label shown in the sidebar switcher and in email notifications. Only the owner can rename.
Agent Controls
Control when and how the agent operates
Pause Agent
Finishes the current task, then stops. Queued triggers won't fire until you resume.
Emergency Stop
Cancels all pending events, disables triggers, and halts the agent immediately. The nuclear option — use when something needs to stop right now.
Team
Invite people, assign roles, and see who has access to what.
Team members are listed with their email, role badge, and status. Invite new members by email with either the Member or Admin role. Invites generate a link that expires automatically.
| Permission | Member | Admin |
|---|---|---|
| Use Solmyr & chat | ✓ | ✓ |
| View team members | ✓ | ✓ |
| Manage notifications | — | ✓ |
| Invite & remove people | — | ✓ |
| Manage billing | — | ✓ |
| Manage integrations | — | ✓ |
The Owner has all admin permissions plus the ability to rename, delete, and transfer the company.
Notifications
Email notifications with two layers: company defaults (what new members inherit) and personal overrides (what you actually receive). Admins can push defaults to all members at once.
Daily Brief
AI-generated project summary — completed tasks, pending reviews, decisions. Configurable frequency (daily or weekly), time, and day.
Low Credits
Fires when your credit balance drops below a threshold you set. Useful for avoiding surprise pauses mid-work.
Decision Needed
Triggered for tactical and strategic decisions. Operational decisions are silent — nobody needs an email about a formatting choice.
Task Ready for Review
When Solmyr finishes a task and moves it to the review column. Your cue to check the work and approve or send it back.
Billing
Two ways to pay. Pick the one that fits your setup.
Bring Your Own Key
Plug in your Anthropic or OpenAI API key. Costs go directly to your provider account. You get budget controls — daily and monthly limits in USD — with optional auto-pause when limits are hit.
Key is stored encrypted and never shared.
Platform Credits
Buy credits through Stripe, we handle the rest. Tiers from $5 to $200 with token estimates. Credits are added instantly. Choose between single-payer (one shared payment profile) or per-member billing.
Recent transactions and balance visible on this tab.
Both modes include usage analytics — token consumption over time, daily and monthly totals, and a budget-remaining indicator so you always know where you stand.
AI Model
Choose your preferred provider and model. This controls which LLM powers Solmyr's reasoning, planning, and tool use.
Anthropic
Claude Sonnet 4.6
Claude Opus 4.6
Claude Haiku 4.5
OpenAI
GPT-5.2
GPT-5 Mini
GPT-4.1
Integrations
Connect external tools so Solmyr can read and act on your data. Two integration types cover different needs.
MCP Servers
Model Context Protocol — the standard way to give LLMs access to external tools. Point Solmyr at a streamable HTTP endpoint and it discovers available tools automatically. Optional auth token support.
Quick-add templates for common services.
REST Connectors
Bridge any REST API. Define a base URL, optional auth, and a list of endpoints with names, methods, paths, and parameters. Each endpoint becomes a tool the AI can call — turning any API into an AI-usable integration without writing code.
Supports path parameters, query parameters, and request bodies.
Activity
A chronological feed of every action, update, and decision across your company. Filter by type — Actions, Updates, Decisions, or Errors — to find what you're looking for. Paginated with a "Load more" button for older entries.
This is the same data that powers the dashboard activity feed, but unfiltered and with full history. Useful for auditing, debugging, or just understanding what the AI has been doing while you were away.
Danger Zone
The bottom of the General tab has a company deletion option — owner only. Permanently deletes all data including workspaces, decisions, tasks, files, messages, and activity history. If you have credits, you'll get the option to transfer them to another company before deletion. This cannot be undone.
That's all of them
Back to Features
Return to the feature overview to explore any module you missed.