Platform Features
Everything you need for trustworthy AI agency. New to the terminology? See ourGlossary.
Core Platform
Solace Browser
Custom browser with scoped permission management, recipe execution, and activity logging. Three surfaces: web automation, machine access, and cloud tunnel.
Solace CLI
Terminal-first orchestration. Manage skills, run recipes, and coordinate agents from the command line. Extends Stillwater CLI with a secure permissions vault and cloud sync.
App Store
Curated collection of verified AI agent workflows. Gmail triage, Slack digest, LinkedIn outreach, and more. Each app tested to rung 65537 before shipping.
Governance
Scoped Permissions
Fine-grained controls:gmail.read.inbox, machine.execute.shell, tunnel.connect. Every permission is time-bounded, revocable, and auditable.
Budget Enforcement
Action budgets (emails/day, posts/day), spend budgets (transaction caps, daily limits), and recipient allowlists. Fail-closed if any gate fails.
Step-Up Approval
High-risk actions (delete, send, spend) always require human sign-off. Preview → Cooldown → Approval → Execution. No shortcuts.
Evidence System
Hash-Chained Audit Trails
Every state transition produces a SHA-256 hash linking to the previous event. Tamper-evident by construction. Append-only activity logs.
Three Evidence Modes
Mode A:Screenshots + action log.Mode B:Full page archives + DOM.Mode C:Full archives + PZip compression (Part 11 Architected).
FDA Part 11 Architected
Sealed approvals, edit tracking with override reasons, electronic signatures, and 90-day retention. Built for regulated industries (customer validation required for full certification).
Intelligence
Recipe System
First run uses LLM to discover the workflow. Subsequent runs replay the recipe at ~69% lower cost (at 70% hit rate). Recipes accumulate as platform capital.
L1–L5 Abstraction
Users buy outcomes, not models. L1 (fast) through L5 (adversarial). Routable across providers. Swap engines without user disruption.
Yinyang Assistant
Page-aware chat helper. Summarizes pages, suggests clicks, builds recipes, and tracks credits. Available on app store, browser, and documentation pages.
Quality & Global Reach
Built-In Quality Checks
Every app and recipe is tested against 47 quality dimensions — covering design, security, accessibility, compliance, and performance — before it reaches you.
47 Languages, Zero CDN
Solace Browser and solaceagi.com are fully translated into 47 languages. All translations are self-hosted — no CDN dependency, no data leaving your machine.
Persistent Conversation Threading
Conversations are trees, not lists. Every thread branches, persists across sessions, and can be resumed from any node. Your AI remembers the full reasoning branch, not just the last message.
Pricing
Built for AI Agents Too
Solace is the first browser automation platform built with AI coding assistants as first-class users. Every capability is available via API, documented for agents, and evidence-captured by default.
Agent Instructions
Download CLAUDE.md snippets, .cursorrules, and Copilot instructions from/agents. One curl command adds Solace to any project.
Machine-Readable Capabilities
/agents.jsondocuments every API endpoint, OAuth3 scope, pricing tier, and hello-world example in structured JSON. Any orchestration framework can consume it.
MCP Server (Coming Soon)
A Model Context Protocol server will let Claude Code and other MCP-compatible tools callsolace:navigate() and solace:screenshot()as native tool calls — no curl required.
QA-as-a-Platform
The first browser that tests itself — and lets you test anything. 7 QA app types, all free, all evidence-chained.
Visual QA
Screenshot comparison, CSS token compliance, responsive breakpoint testing. Automated via localhost:8888 API.
API QA
Endpoint smoke tests, response schema validation, latency monitoring. Every result hash-chained.
Accessibility QA
WCAG 2.1 AA compliance, ARIA audit, keyboard navigation, focus order. Built into every page load.
Security QA
XSS probing, CSRF checks, injection tests, CORS policy, CSP headers. Continuous scanning.
Performance QA
Core Web Vitals (FCP, LCP, CLS), memory profiling, load optimization. Real browser metrics.
Evidence QA
Hash chain integrity, ALCOA+ compliance, FDA Part 11 audit readiness. Tamper-evident proof.
Integration QA
Cross-app flows, domain-to-domain handoff, end-to-end user journeys. Full system verification.
AI coding agents (Claude Code, Codex, Cursor) can extend QA apps to test any web project. Not just CLI — the entire web.