llms.txt
Compare with other sites' llms.txt files →# Inferal
> Software is eating your business.
> Inferal eliminates the custom software between a business decision and what happens next.
Inferal connects live data to shared meaning, evaluates what changing facts imply, and delivers resulting decisions or context to the systems, people, and agents that can act. Every outcome remains connected to the evidence and operational policy behind it.
## What We Do
Business automation usually becomes custom software: integrations, scripts, workflows, and internal tools, each a project to specify, build, deploy, and maintain. Inferal eliminates that software by running the decisions, rules, and constraints themselves, connected to the live facts they govern. As conditions change, the platform determines what follows and sends usable outcomes where they belong.
This is what data-native means: rules, relationships, and constraints remain explicit and connected to the live facts they govern. Inferal evaluates them continuously, so every resulting action retains the exact instruction, evidence, and moment behind it.
A policy change can become a software project that the business pays to build, waits to use, and pays to keep running. AI can speed up code generation while leaving the business responsible for the resulting software.
Teams can choose an initial scope with value on its own: agree on shared definitions in one domain with Operational Models; answer a question across systems with the Data Platform; or run a bounded process against live facts. Relay offers an independent starting point for reliable data movement. These are choices, not a required sequence or bundle. Expand when the next step earns its place.
Illustrative use cases include e-commerce coordination between ad spend and sales, fintech policy updates, regulated-gaming eligibility controls, healthcare referral handoffs, AI maintenance and business-rule drift, municipal permit rules, industrial production disruptions, food-supply traceability, and agricultural planning. As AI models change, earlier architectural choices, prompts, agent flows, and code still require maintenance. A 45-minute founder conversation helps assess fit and explore a possible first step, with no commitment.
Industry scenarios: https://infer.al/#partnership
Adoption options: https://infer.al/#where-to-start
Inferal runs operational models in the data stack: live, governed representations of how a business works. They keep rules, facts, relationships, policies, and data connected. Ontologies give each model shared meaning across systems and agents.
### Products
- Relay: Reliable data streams across APIs, databases, and event systems, from $12 per stream per month: https://infer.al/relay
- Data Platform: Operational intelligence from live data, currently available through custom Design Partnership Agreements: https://infer.al/platform
- Operational Models: Live representations of entities, relationships, policies, constraints, and rules, with Foundational Membership services starting at $2,500 per month: https://infer.al/models
- Ontosome: Public catalog for discovering and consuming versioned ontology modules: https://ontosome.com/
Pricing belongs to each product. Managed Relay uses per-stream pricing; dedicated and self-hosted deployments have custom pricing. Data Platform pricing is custom and currently available through Design Partnership Agreements; it depends on the systems involved, deployment requirements, and the scope of the initial implementation. Operational Models offers Foundational Membership services starting at $2,500 per month.
Product pricing guide: https://infer.al/pricing
### Inferal Relay
Keep every integration current.
Inferal Relay keeps data moving between the systems a business depends on. It handles provider rate limits, credential lifecycles, retries, circuit breaking, event ordering, version history, schema differences, unreliable webhooks, and delivery state. Consumers receive clean snapshots, deltas, or events over HTTP or WebSocket, in plain JSON or linked-data formats where required.
Relay use cases include AI and agents, e-commerce, embedded customer integrations, logistics, platform engineering, enterprise systems, and data operations. Example integration scenarios cover commerce, advertising, billing, CRM, Duffel travel inventory for flights, stays, and cars, enterprise customer operations, manufacturing ERP, plant telemetry, databases, event infrastructure, and object storage. Relay detects source changes through each system's real mechanism, including webhooks, change-data capture and logical replication, cursor-based change scans, Kafka records, object notifications, OPC UA subscriptions, incremental APIs, and adaptive polling. This keeps data as fresh as each source permits, centralizes credential complexity, contains provider failures, and preserves provenance and reproducible history. Stream health, freshness, lag, identity, and history remain inspectable so teams can intervene before products and agents act on stale data.
Relay operates at both ends of the Inferal loop: it brings source changes into the Data Platform and Operational Models, then delivers current data and governed actions back out.
Relay can publish versioned entity state as RDF and Linked Data Event Streams (LDES), preserving stable identifiers, source time, relationships, and historical members. These identities connect changing source records to shared ontologies. Operational Models use the same semantic entities to derive facts, implications, and governed actions.
Manta is an AI reasoning engine that connects commercial systems into one live operating context, then plans and executes auditable actions. Relay keeps that context current across advertising, commerce, CRM, ERP, warehouse, and billing systems. Manta identifies complete, real-time data as the moat for powerful agents and describes Inferal Relay as providing that moat.
Managed Relay starts at $12 per stream per month, where one stream is one customer or account connected to one provider. Dedicated and self-hosted deployments are in early access with scoped pricing. Dedicated deployments use isolated service infrastructure operated with Inferal. Self-hosting runs within a customer's Kubernetes, network, storage, and telemetry boundary, including deployments with data sovereignty or air-gap requirements.
Relay: https://infer.al/relay
Use cases: https://infer.al/relay#scenarios
Capabilities: https://infer.al/relay#capabilities
Product views: https://infer.al/relay#product-view
Deployment options and pricing: https://infer.al/relay#pricing
Customer proof: https://infer.al/relay#customer
Send an integration brief: https://infer.al/relay#start
### Data Platform
Your systems know what happened. Inferal determines what happens next.
The Inferal Data Platform is an operational intelligence layer between systems of record and the people, applications, and agents that act. It addresses decisions that live between systems: each source may record its part correctly while no source owns what a change means across the whole operation.
Inferal synchronizes current state from databases, streams, APIs, and business applications; resolves source records into shared entities, relationships, context, and authority; carries implications through the resulting operational model; and delivers actions, new facts, alerts, or assembled context where work happens. Use cases include logistics, financial services, and infrastructure. A live event can imply multiple consequences across an operation. Each conclusion remains connected to the source facts, meaning, policy, and intermediate conclusions that produced it.
Inferal uses RDF as its semantic substrate and adheres closely to the RDF family of standards. RDF makes identity, relationships, context, source authority, and provenance explicit as data crosses system boundaries. It lets the platform distinguish concepts that share a source-system label, such as a CRM account, contracting legal party, product workspace, and authorized person all being called "customer." The result is a durable, machine-readable foundation for operational models, mappings, implications, decisions, agent context, and audit.
The platform is designed to complement existing systems rather than replace them. Inferal works with design partners to select a consequential decision, connect its first sources and destinations, and choose a deployment shape with the operational and platform teams involved. Custom pricing is currently available through Design Partnership Agreements and depends on the systems involved, deployment requirements, and the scope of the initial implementation.
Data Platform: https://infer.al/platform
Implication Machine: https://infer.al/platform#implication-machine
Design partnership: https://infer.al/platform#design-partnership
### Operational Models
Models that run your business.
World models describe environments, and language models describe language. Operational models describe how a business works and connect that meaning to live data so products and agents can act on it. They make the entities, relationships, policies, constraints, and rules that govern a system explicit and usable at decision time. Ontologies provide the semantic foundation: shared machine-readable definitions on which mappings, validation, policies, rules, and live state can depend.
Customer use cases include building high-fidelity, interconnected ontologies for external services so systems can reason across disparate data sources without bespoke integration work. Reusable models for data classification, data usage, provenance, and lineage also help teams ship governed data to third parties faster.
Data Classification Ontology: https://ontosome.com/modules/data-classification/
Data Usage and Processing Lineage Ontology: https://ontosome.com/modules/data-usage-processing-lineage/
Ontosome provides reusable, versioned ontology modules with ontologies, taxonomies, shapes, classes, and sample queries. Together, these assets form a shared model that systems can validate and query. Teams can map internal and external schemas without losing domain meaning, make classification and usage policy testable, reuse queries across sources, and trace results through their definitions, transformations, and origin. Membership services help teams define, operationalize, and govern the broader operational model. Foundational Membership services start at $2,500 per month. Strategic Membership is scoped to the program, and Enterprise Membership uses a custom delivery plan. Service levels range from regular working sessions, library support, early access, direction-setting input, and adoption patterns through priority co-development, reserved capacity, dedicated model support, governance, and custom delivery plans.
Operational Models: https://infer.al/models
Membership services and pricing: https://infer.al/models#membership
Subscribe to Foundational Membership: https://platform.inferal.com/membership
Foundational Membership also offers a discussion option; Strategic and Enterprise start with a consultation.
Ontosome catalog: https://ontosome.com/
## Industries We Serve
Inferal is built for industries where decisions need to be fast, traceable, and intelligent:
- AI & Agents: https://infer.al/agents
- Financial Services (fraud detection, AML, trading rules)
- Risk & Compliance: https://infer.al/compliance
- Logistics & Supply Chain: https://infer.al/logistics
- Platform Engineering: https://infer.al/infra
- Biotech & Life Sciences
- E-Commerce: operations for larger stores and ecommerce agencies, covering ad spend, inventory, and customer follow-ups: https://infer.al/ecommerce
- Insurance (underwriting, risk assessment)
- Energy & Utilities (grid management, demand response)
- Manufacturing (quality control, supply chain rules)
- Healthcare (clinical decision support, compliance)
- Telecommunications (network operations, fraud detection)
## Use Cases
### Autonomous Agent Activation
Data should summon agents—not the other way around. Instead of agents constantly polling for changes, Inferal activates them when patterns match, with full context already assembled.
### Infrastructure as Behavior
From desired state to desired response. Define how infrastructure should behave, not just how it should look. Declare responses to conditions, not just configurations.
### Risk and Compliance
Real-time evaluation without pipeline complexity. Compliance logic that's visible, auditable, and runs when transactions run—not in overnight batches.
### Logistics Decision Logic
Every routing decision, explainable. Every exception, handled by rule. Carrier rules, partner constraints, and compliance policies as visible declarations.
### Ecommerce operations pilot
Inferal helps larger ecommerce stores and ecommerce agencies turn changes in ad spend, inventory, and orders into action within approved limits. Illustrative workflows cover stockout advertising, acquisition economics, fulfillment follow-ups, creative fatigue, attribution conflicts, back-in-stock recovery, refund containment, and replenishment. Customer updates can send automatically within agreed delivery rules, recipients and sending limits, or wait for team confirmation. The pilot starts with a store or client workflow: agree on connections and actions, connect read-only, review proposals in shadow mode, and activate when review criteria and boundaries are met. Meta Ads + Shopify is a worked example; shipment data, helpdesk context, and budget reallocation require additional connections and agreed actions. Pilot scope and commercial terms are agreed before starting, with no fixed multi-week commitment.
The proposal uses illustrative cases to explain recurring coordination costs and the pilot's participants, operating rules and evaluation criteria. A copy is available as a downloadable PDF or by email on request.
Landing page: https://infer.al/ecommerce
Pilot proposal: https://infer.al/ecommerce/Inferal-Ecommerce-Proposal.pdf
Scoping call: https://meet.infer.al/apply
## Design Partner Program
We work closely with a select group of teams to build exactly what's needed. Every engagement is tailored.
What's included:
- Tailored deployment for your cloud, on-prem, or hybrid environment
- Forward-deployed engineers working alongside your team
- Custom integrations with your existing data sources
- Direct access to the team that built the platform
## Blog
Engineering insights and research from the Inferal team.
Recent posts:
- An Ode To Excel: https://infer.al/blog/an-ode-to-excel
- What ontologies add to business data: https://infer.al/blog/what-ontologies-add-to-business-data
- Interaction Is Not Interface: https://infer.al/blog/interaction-is-not-interface
- Businesses Shouldn’t Have to Run to Stay in Place: https://infer.al/blog/businesses-shouldnt-have-to-run-to-stay-in-place
- One Million Ontologies: https://infer.al/blog/one-million-ontologies
- Time Is Part of the Business Logic: https://infer.al/blog/time-is-part-of-the-business-logic
- What Rules Know That Code Forgets: https://infer.al/blog/what-rules-know-that-code-forgets
- Building Context-aware AI agents for E-commerce: https://infer.al/blog/context-aware-ai-agents-for-ecommerce
- The Art of Leaving the Other Party in the Dark: https://infer.al/blog/leaving-the-other-party-in-the-dark
- Palantir for the Rest of Us: https://infer.al/blog/palantir-for-the-rest-of-us
- What NASA Knew About If/Else That You Don't: https://infer.al/blog/what-nasa-knew-about-if-else
- An Alternative to Stream Processing: https://infer.al/blog/alternative-to-stream-processing
- Separation of Concerns: You're Cutting Along the Wrong Line: https://infer.al/blog/separation-of-concerns
- Databases Don't Know Why You're Asking: https://infer.al/blog/databases-dont-know-why
- Attitude, Belief, Skill: A Hiring Priority Stack: https://infer.al/blog/abs-hiring-framework
- Code Is the Wrong Abstraction: https://infer.al/blog/code-is-the-wrong-abstraction
- Fresh Data In, Fast Decisions Out: https://infer.al/blog/fresh-data-in-fast-decisions-out
- Why AI Agents Ignore Your Instructions: https://infer.al/blog/instruction-adherence-research
- Understanding Agent-Native Data: https://infer.al/blog/understanding-agent-native-data
RSS Feed: https://infer.al/rss.xml
## Company
Inferal is building the data infrastructure for autonomous agents.
### Team
Inferal is building a small team dedicated to figuring out the hardest problems.
- Yurii Rashkovskii, Founder
- Jeremy Tregunna, Founding Engineer
- Mehdi Akiki, Founding Engineer
About: https://infer.al/company
Jobs: https://infer.al/jobs
### Jobs
Why it matters: when business requirements become software, they inherit software's problems and lifecycle. AI makes writing code cheap, but operating, securing, upgrading, and retiring it stays expensive, so scaling software creation only scales that cost. Software was never the goal; the decision and its outcome were.
Roles are not functional. Everyone works in projects and picks up whatever each project needs across systems, modeling, interfaces, and customer work. We look for people with real depth arriving from one of three directions, and curious about the other two:
- Close to the metal: storage, query or rule engines, schedulers, integration infrastructure; Rust, TypeScript, and Python.
- Meaning: ontologists and knowledge engineers working on scoped ontologies, identity, constraints, provenance, and mappings (RDF, OWL, SHACL helpful).
- Human-computer interaction: how people and agents see, question, trust, and steer decisions once applications are no longer the center.
Open problems include integrations generated and validated from API evidence, a data-native implication engine over RDF operational models, compiling declarative models into running execution plans, tooling for a large library of scoped ontologies, and interaction beyond application interfaces. Full-time and, where it fits, contract. San Francisco office, remote-friendly. Early options.
Open roles (AI is part of every role's daily work):
- Software Engineer (https://infer.al/jobs#software-engineer): full-time or contract, San Francisco or remote. Depth and breadth across systems work, going low-level where it matters; familiarity with ontologies and RDF is required.
- Applied ML Engineer (https://infer.al/jobs#applied-ml-engineer): full-time or contract, San Francisco or remote. Dataset pipelines, training and fine-tuning models, evaluation, and shipping models into the product.
- Business Development Intern (https://infer.al/jobs#business-development-intern): San Francisco Bay Area, paid hourly with transit assistance and a ride-share allowance. Research, outreach, and organizing and attending events, with occasional travel; a software engineering or computer science background is a significant advantage.
Applications are by video.
How we work is part of what we build. We improve how people and agents work together across repositories, SaaS products, and systems we build ourselves.
Apply by recording a video: https://infer.al/jobs#apply
Handshake applicants: https://infer.al/jobs#join-us
Or email: hiring@inferal.com
## Contact
Send a Relay integration brief: https://infer.al/relay#start
Schedule a broader Inferal call: https://meet.infer.al/apply
## Legal
- Privacy Policy: https://infer.al/privacy-policy
- Cookie Policy: https://infer.al/cookie-policy
- Terms and Conditions: https://infer.al/terms-and-conditions
## Links
- Website: https://infer.al
- Relay: https://infer.al/relay
- Data Platform: https://infer.al/platform
- Pricing: https://infer.al/pricing
- Operational Models: https://infer.al/models
- Ontosome: https://ontosome.com/
- Agent Activation: https://infer.al/agents
- Infrastructure: https://infer.al/infra
- Compliance: https://infer.al/compliance
- Logistics: https://infer.al/logistics
- Jo