Skip to content
SLM-Works

Our models

SLM-Works Models

Task-specific SLMs we built and fine-tuned for real business workflows. Each model runs on your infrastructure, delivers up to 96% token savings versus a general-purpose LLM, and keeps your data private.

Available now

7 models
Available

SLM-Works Invoice Generator

Structured invoice creation from unstructured input

Fine-tuned to extract line items, tax calculations, payment terms, and vendor details from emails, PDFs, and chat messages - then output clean, structured invoice data ready for your ERP or accounting system.

3B~92% fewer tokens per invoice vs GPT-4
  • Extract line items from unstructured text, email, or scanned PDFs
  • Multi-currency support with automatic tax calculation logic
  • Output structured JSON matching common ERP schemas
  • Handles partial or messy input gracefully with confidence scores

Ideal for

Finance teamsAccounting firmsERP integratorsProcurement departments
Available

SLM-Works Contract Analyzer

Clause extraction and risk flagging for legal documents

Purpose-built to read contracts, extract key clauses (liability, termination, IP, SLA), flag deviations from your standard templates, and surface risk areas - without sending confidential documents to external APIs.

3.8B~88% fewer tokens per contract review vs GPT-4
  • Extract and classify 20+ standard contract clause types
  • Flag deviations from your baseline templates
  • Risk scoring per clause with plain-language explanations
  • Works entirely on-premise. zero data egress

Ideal for

Legal departmentsProcurement teamsCompliance officersLaw firms
Available

SLM-Works Support Classifier

Instant ticket routing and intent detection

Trained on enterprise support patterns to classify incoming tickets by category, urgency, sentiment, and required skill group - enabling instant routing and SLA-aware prioritization without manual triage.

1B~96% fewer tokens per classification vs GPT-4
  • Classify tickets across 50+ configurable categories
  • Detect urgency level and customer sentiment in one pass
  • Sub-10ms inference latency for real-time routing
  • Integrates with ServiceNow, Jira, Zendesk, and custom APIs

Ideal for

Customer support teamsIT service desksManaged service providers
Available

SLM-Works Doc Summarizer

Concise summaries of long-form business documents

Fine-tuned for enterprise document types. board reports, policy documents, technical specs, RFPs. to produce accurate, structured summaries with key decisions, action items, and risk callouts highlighted.

2B~90% fewer tokens per summary vs GPT-4
  • Summarize documents up to 32K tokens in a single pass
  • Extract key decisions, action items, and deadlines
  • Configurable output: executive brief, bullet points, or structured JSON
  • Preserves technical terminology and proper nouns accurately

Ideal for

Executive assistantsBoard secretariesAnalystsProject managers
Available

SLM-Works Email Drafter

Professional emails from bullet points or voice notes

Fine-tuned to transform brief instructions, bullet points, or raw voice transcripts into well-structured, on-brand professional emails. Adapts tone and formality to recipient context and produces send-ready drafts in seconds.

2B~91% fewer tokens per draft vs GPT-4
  • Draft complete emails from 3–5 bullet points or a raw voice note
  • Tone adaptation: formal, consultative, assertive, or friendly
  • Context-aware subject line and opening line generation
  • Runs inside your mail client or CRM with no external API calls

Ideal for

Sales teamsAccount managersExecutive assistantsCustomer success
Available

SLM-Works HR Screener

CV screening and candidate data extraction at scale

Purpose-built to parse CVs and cover letters, extract structured candidate profiles, and score applicants against job requirements you define. Reduces first-pass review time from hours to seconds without sending sensitive data outside your network.

1B~95% fewer tokens per CV vs GPT-4
  • Extract 25+ structured fields: skills, tenure, education, certifications
  • Score candidates against weighted criteria from your job description
  • Flag missing requirements and experience gaps with plain explanations
  • Outputs structured JSON ready for your ATS or HR system

Ideal for

HR teamsTalent acquisitionStaffing agenciesShared services centers
Available

SLM-Works Feedback Analyzer

Structured insight from customer feedback and NPS responses

Classifies and extracts themes, sentiment, urgency, and product signals from free-text survey responses, NPS comments, support reviews, and social feedback. Turns unstructured voice-of-customer data into consistent, queryable output.

2B~92% fewer tokens per response vs GPT-4
  • Multi-label sentiment and topic classification in one pass
  • Extract product mentions, feature requests, and churn signals
  • Urgency and escalation scoring for negative feedback
  • Batch mode for processing thousands of responses per run

Ideal for

CX teamsProduct managersMarket researchBrand teams

Coming soon

5 models
Coming soon

SLM-Works Data Extractor

Structured data from messy sources

Purpose-built to pull structured data from emails, forms, reports, and semi-structured documents. names, dates, amounts, reference numbers. and output clean, validated records for downstream systems.

3B~91% fewer tokens per extraction vs GPT-4
  • Entity extraction across 30+ field types (dates, amounts, names, IDs)
  • Handle multi-language input with consistent output schema
  • Built-in validation and confidence scoring per field
  • Batch processing mode for high-volume pipelines

Ideal for

Data engineering teamsOperationsInsurance claimsBanking back-office
Coming soon

SLM-Works Code Reviewer

Automated code review for common enterprise stacks

Fine-tuned on enterprise code review patterns to flag security issues, style violations, performance bottlenecks, and logic errors. runs entirely in your CI/CD pipeline without external API dependencies.

6.7B~85% fewer tokens per review vs GPT-4
  • Security vulnerability detection (OWASP Top 10 patterns)
  • Style and convention checks for Java, Python, TypeScript, Go
  • Performance anti-pattern flagging
  • Inline suggestion format compatible with GitHub PR reviews

Ideal for

Engineering teamsDevOpsSecurity teamsPlatform engineering
Coming soon

SLM-Works Meeting Scribe

Transcripts to structured minutes and action items

Converts raw meeting transcripts into formatted minutes with agenda sections, key decisions, action items with owners, and deadlines. Works with transcripts from any video conferencing tool and adapts to your minute template.

3B~89% fewer tokens per meeting vs GPT-4
  • Detect and extract action items with assigned owner and deadline
  • Segment transcript by agenda topic automatically
  • Configurable output: executive summary, full minutes, or action list only
  • Handles cross-talk and disfluencies without manual cleanup

Ideal for

Operations teamsProject managersBoard secretariesConsulting firms
Coming soon

SLM-Works RFP Analyzer

Tender and RFP parsing for faster bid decisions

Parses RFPs, tenders, and procurement documents to extract requirements, evaluation criteria, mandatory qualifications, deadlines, and scoring weights. Surfaces bid/no-bid signals so your team focuses effort on winnable opportunities.

3.8B~87% fewer tokens per RFP vs GPT-4
  • Extract mandatory and scored requirements with page references
  • Identify evaluation criteria and their weighting from tender documents
  • Flag compliance risks, missing prerequisites, and tight deadlines
  • Structured output for bid management tools and CRM deal records

Ideal for

Business developmentProcurement teamsConsulting firmsGovernment contractors
Coming soon

SLM-Works Compliance Checker

Policy conformance screening for outgoing content

Screens outgoing communications, contracts, marketing content, and internal documents against your defined policy rules. Flags violations, suggests compliant rewrites, and logs every check for audit purposes - entirely on your infrastructure.

3B~90% fewer tokens per check vs GPT-4
  • Check content against 100+ configurable policy rules simultaneously
  • Violation explanation with specific clause reference and severity
  • Suggest minimal compliant rewrites preserving original intent
  • Full audit log of every check, version, and outcome

Ideal for

Compliance teamsFinancial institutionsHealthcare operatorsMarketing legal review

Need a model fine-tuned for a different workflow?

We scope and build custom SLMs for specific document types, tasks, or industry vocabularies. Typical engagement starts with a PoC milestone.