Map the real workflow
Capture the trigger, systems, people, decisions, exceptions, and handoffs that actually run the operation.
Early beta · open source · honest maturity labels
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 the execution or integration layer that fits the job.
Operational use cases
Real workflows, with clear implementation boundaries.
Every example states what SolveLang can model today and where external implementation or experimental AI behavior begins.
Support triage
Route urgent and billing cases with readable deterministic policy.
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; durable approval execution remains planned.
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 still 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 pretending to execute 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 workflow analysis and a smaller browser-safe runtime subset.
4. Future adapters
External runtimes and managed execution remain evidence-led roadmap items.
Product status
Know what you are looking at.
- • Rust CLI with lexer, parser, AST runtime, validation, diagnostics, imports, arrays, objects, functions, loops, and JSON helpers
- • Hardened local execution modes for restricting network, file, environment, and AI capabilities
- • Local-first Workflow Intelligence Studio with deterministic analysis and simulation
- • Browser-safe preview for a deliberately smaller syntax subset
- • HTTP, file, and environment helpers
- • AI agent syntax with local fallback and optional OpenAI-backed responses
- • Test-mode API access, account, subscription, and metering infrastructure
- • Stable language specification and release packaging
- • Broader provider and execution adapters
- • Production integrations and managed workflow execution
- • Team environments, hosted observability, and production operating controls
For teams
Need help making a workflow understandable?
The near-term business path is service-led: workflow clarity audits, prototype sprints, and implementation in the right execution stack for the client.
Discuss a workflow auditFor recruiters + developers
Inspect the engineering decisions, not just the marketing page.
The repository demonstrates language implementation, Rust systems work, TypeScript product engineering, AWS/serverless infrastructure, IAM, testing, AI workflow design, safety boundaries, and technical product strategy.
Open GitHub repositoryFAQ
The important boundaries, answered directly.
What is SolveLang?
SolveLang is an early-beta readable, explainable workflow language designed for AI-assisted business processes. Its canonical runtime is a Rust CLI.
Is SolveLang production ready?
No. SolveLang is an early-beta project. The Rust CLI works locally, while hosted production execution and broader production integrations remain planned.
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. The browser preview intentionally supports a smaller safe subset. The Rust CLI remains the canonical validator and runtime.
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?
Yes as a design and implementation layer. A SolveLang workflow can be used to clarify what should happen before implementing the final execution in an existing platform or custom service. Managed adapters are still roadmap work.