Map the real workflow
Capture the trigger, systems, people, decisions, exceptions, and handoffs that actually run the operation.
Open source · explicit capability boundaries
See the system before you automate it.
A readable, explainable workflow language for AI-assisted business processes.
Make deterministic rules, AI-assisted decisions, approvals, tools, failure paths, and human review points understandable before a process becomes opaque automation.
The Rust CLI is the canonical runtime. Studio analysis is deterministic. The browser preview supports a smaller safe subset. Managed production execution is planned, not available today.
support_triage.solvecanonical syntax
let ticket = {
topic: "billing",
priority: "urgent"
}
if ticket.priority == "urgent" {
print("Action: escalate to founder today")
}
if ticket.topic == "billing" {
print("Owner: finance operations")
}Workflow X-Ray
Trigger
Customer support ticket
Decision
Urgent or normal priority?
Human review
Escalate urgent cases
Owner
Finance operations for billing
Designed for the workflows teams repeat every day.
SupportClient intakeLead routingApprovalsInternal reporting
How it works
Understand the workflow before choosing the execution layer.
SolveLang starts with decisions, ownership, exceptions, and review points—not with a connector canvas.
Make every branch reviewable
Turn hidden judgment calls into readable rules with explicit ownership, failure paths, approvals, and human review.
Automate with control
Validate the workflow with the canonical Rust runtime, then implement it using an explicitly authorized execution or integration layer.
Operational use cases
Real workflows, with clear implementation boundaries.
Every example distinguishes modeling from live execution and provider-dependent behavior.
Support triage
Explore classification, ownership, and review rules without sending a message or creating a remote task.
Lead qualification
Make fit, urgency, and follow-up rules visible before CRM automation.
Client intake
Model missing-information checks, ownership, and handoffs before implementation.
Invoice processing
Model routing and checks today; extraction and accounting integrations remain implementation work.
Approval workflows
Model approval policy today; general managed approval execution is not claimed.
Email summarization
Use optional AI-assisted behavior with explicit review rather than silent business action.
Workflow Intelligence Studio
Explore the workflow before writing automation.
Studio is a local-first workspace for workflow modeling, deterministic analysis, simulation, traces, versions, quality analytics, and evidence. Its analysis is not AI-generated.
- Local-first workflow data and analytics
- Deterministic checks and scenario simulation
- Workflow graph, traces, versions, and evidence
- Generated .solve drafts that require canonical Rust validation
Local-first model
Graph
See triggers, rules, handoffs, and review points.
Analysis
Run deterministic checks across modeled workflow structure.
Simulation
Explore defined scenarios without executing production systems.
Evidence
Keep traces, versions, and Workflow X-Ray findings reviewable.
Architecture
One readable workflow layer, multiple ways to explore and run it.
1. Language
Lexer, parser, AST, validation, diagnostics, and runtime semantics in Rust.
2. Local execution
Trusted local runs plus hardened modes that deny sensitive capabilities.
3. Studio + preview
Local-first deterministic analysis and a pinned safe-core browser preview.
4. Qualified integrations
Live provider connections and managed execution require separate implementation and evidence.
Product status
Know what you are looking at.
Capability evidence reviewed 2026-09-17. These are implementation and deployment labels, not live service-health measurements.
Working locally
Implemented tools you can use without connected production automation.
- Rust CLI with parsing, validation, diagnostics, imports, data structures, functions, loops, and JSON helpers
- Hardened local execution policies for network, file, environment, and AI capabilities
- Local-first Workflow Intelligence Studio with deterministic analysis and simulation
Browser previews
The workflow audit and support triage are local, rule-based planning previews. They do not connect an inbox, call an AI model, create remote tasks, or send messages.
- Pinned WebAssembly build of the canonical safe core with host side effects denied
- Automatic workflow mapping and support triage examples; proposed actions are not executed
Deployed, with gates
Production API access, customer-account infrastructure, authenticator 2FA, and Admin CRM were verified enabled on September 17, 2026. This is deployed infrastructure, not a claim of general managed workflow execution or an availability SLA.
- Public site on AWS Amplify
- Production API/customer-account and authenticator infrastructure; individual features remain gated
- No independently measured uptime or SLA is published
Experimental
Implementation and tests do not by themselves qualify a provider for production use.
- Provider-backed AI and side-effecting HTTP, file, and environment helpers
- Provider adapters and Self-Driving activation require scoped credentials and explicit authorization
Billing enabled; payment canary pending
Production API subscription billing, checkout, Stripe price validation, webhook signature enforcement, and billing-failure monitoring were verified enabled on September 17, 2026. The first real-payment canary is still pending, so broader public payment availability is not yet claimed.
- Developer, Pro, and Business subscription checkout and recurring billing are enabled
- Paid priority and provider activation remain separate rollout gates
Remaining launch work
Not represented as complete or generally available.
- Connected inbox-to-action support automation
- General managed workflow execution and broader qualified integrations
- Additional platform evidence, independent monitoring, and operational exercises
For teams
Make the next handoff understandable.
Describe the process directly in the workflow audit preview. Inspect the proposed map before connecting any production integration.
Map a workflow nowFor recruiters + developers
Inspect the engineering decisions, not just the marketing page.
Explore the language implementation, TypeScript tooling, AWS infrastructure, IAM, testing, and explicit safety boundaries in the source repository.
Open GitHub repositoryFAQ
The important boundaries, answered directly.
What is SolveLang?
SolveLang is an open-source, readable, explainable workflow language designed for AI-assisted business processes. Its canonical runtime is a Rust CLI.
Is SolveLang production ready?
The public website is deployed, and the Rust CLI and local analysis tools work today. Production API access, customer-account infrastructure, authenticator 2FA, and Admin CRM were verified enabled on September 17, 2026. This is deployed infrastructure, not a claim of general managed workflow execution or an availability SLA. Production API subscription billing, checkout, Stripe price validation, webhook signature enforcement, and billing-failure monitoring were verified enabled on September 17, 2026. The first real-payment canary is still pending, so broader public payment availability is not yet claimed. General managed workflow execution and connected support automation are not complete.
What is Workflow Intelligence Studio?
Workflow Intelligence Studio is a local-first workspace for modeling workflows, deterministic analysis, simulation, traces, versions, and evidence. Its analysis is deterministic, not AI analysis.
Does the browser preview run the full SolveLang runtime?
No. It uses a pinned WebAssembly build of the canonical safe core with host side effects denied. The native Rust CLI remains canonical for the full language.
How is SolveLang different from automation builders?
SolveLang focuses on making workflow intent, decisions, exceptions, ownership, human review, and failure paths readable before an execution platform or integration layer runs the process.
Can SolveLang work with existing automation tools?
It can clarify and validate the rules before implementation in an existing platform or custom service. Do not interpret a planning preview as an authorized live connection.