Pricing
Transparent about cost.
Honest about what drives it.
We don't publish a fixed price list because AI projects aren't interchangeable. A RAG system for a three-person law firm is a different project than one for a regional hospital. What we can do is show you the ranges we work in, explain what drives the cost up or down, and tell you how we structure engagements so there are no surprises.
Price Ranges
A rough idea of what things cost.
These are honest ranges based on projects we've scoped, not marketing-friendly minimums. Your actual quote will depend on the specifics of your business, your data volume, integration complexity, and the level of customization required.
Don't see what you need? This list isn't exhaustive. If you're looking for professional AI support and your situation doesn't map neatly to one of these categories, reach out anyway — we're likely able to help or point you in the right direction.
We don't operate on a take-it-or-leave-it basis. If you have a budget in mind, tell us. We'd rather scope a meaningful first phase than lose the conversation. Tell us what you want to accomplish and what it's worth to your business — we'll work with you to build a plan that fits.
Software & Integrations
Workflow Automation
Per automation or system
End-to-end automation of business processes — intake, routing, approvals, notifications, and data handoffs between your existing tools. Learn about workflows.
Single basic automation, no ongoing server requirements
from $500
Single linear flow with 1–2 integrations
from $2,500
Multi-workflow system with routing and many integrations
$5,000 – $10,000+
AI Agent Development
Custom autonomous agent
Goal-directed processes that reason, plan, and act across your systems. Can run on your infrastructure, ours, or in the cloud — ongoing API or hosting costs may apply depending on deployment. Learn about agents.
Targeted single-purpose agent (LLM + tools)
from $500
Multi-step agent with integrations, hosted & managed
$5,000 – $10,000+
The terms "workflow automation" and "AI agent" overlap significantly — and the industry uses them interchangeably. What actually drives cost is where the system runs, how many people use it, and how many other systems it connects to. If you're not sure which applies to your situation, that's exactly what the discovery call is for.
Internal or customer-facing
Ingestion pipeline, vector store, retrieval and generation layer, and a user-facing interface. Suited for internal knowledge bases and customer-facing applications alike. Accuracy and reliability scale with how much rigor goes into the build.
Basic: fixed-width chunking, no performance evaluation
from $2,500
Standard: data cleaning, thoughtful chunking strategy, eval testing
$5,000 – $10,000
High-accuracy: eval dataset development, iterative strategy & model selection, optional GraphRAG
$10,000+
Model Fine Tuning
Domain adaptation & specialization
Adapt a pre-trained model to your domain, terminology, or task. Cost is driven by three factors: the size of the base model being fine-tuned, the availability and condition of your training dataset, and your compute budget and required turnaround time.
LoRA / adapter fine-tuning — lightweight behavioral or style adjustments on a smaller model using a well-prepared dataset
$5,000 – $10,000
Full fine-tune — deeper domain adaptation on larger models, data cleaning and curation included, with iterative evaluation
$10,000+
Custom Application Development
Full product build
Full-stack AI-powered application — web or desktop. For clients building a product or internal platform. We scope based on features, not hours. Archivist is an example of this type of project.
Hardware
Entry Inference Server
Single consumer/prosumer GPU
Workstation-class GPU (e.g. RTX 4070–4090 tier) assembled into a quiet, office-friendly tower. Capable of running 7B–30B parameter models comfortably. Budget-conscious builds using quality used components are available. Note: component costs fluctuate with GPU and memory market pricing.
Multi-GPU Rackmount Server
High-throughput rackmount
Multi-GPU rackmount configurations for organizations with high concurrent user demand or large-model requirements. Form factor varies by GPU selection and chassis. Quoted per specification — contact us to scope.
Interchange™ AI Gateway
Mini-PC appliance · Early Access
Compact appliance that brings private AI to any network. Ideal for small offices that want in-house AI without a dedicated server room. Pricing will be announced at general availability.
Professional Services
AI Readiness Evaluation
2 week engagement
An honest look at where your business stands with AI: what you have, what’s missing, and what’s worth doing first. Deliverables include a current-state assessment, prioritized opportunity list, and a written roadmap with sequenced recommendations your team can actually act on.
Independent Technology Review
1–2 week engagement
Buying AI software or hardware from another vendor and want a second opinion before you sign? We evaluate the product on your behalf — examining fit for your environment, data practices, vendor claims, architecture, and total cost of ownership — and give you an unvarnished written assessment.
Team Training
Half-day or full-day workshop
Hands-on, role-specific AI training for your team. Remote or on-site. Priced per session; multi-session programs are scoped separately.
$1,000 half-day
$1,500 full-day
Intelligent Document Scanning
Per-job or high-volume batch
Converts paper records, scanned PDFs, and unstructured files into clean, structured, AI-ready text. Priced by data volume. We also offer wholesale rates for copy shops and scanning services looking to add AI output as an upsell to their customers. Learn more about our scanning service.
from $50
volume pricing available
Ongoing Support
Monthly, rolling
Break/fix, preventive maintenance, model updates, and monitoring. Remote-only plans start lower; plans with included on-site visits are higher.
Cost Factors
What pushes a project toward the high end.
Integration complexity
Connecting to multiple existing systems, legacy APIs, or uncommon data formats takes significantly more time than a greenfield build.
Compliance requirements
HIPAA, GLBA, FERPA, or other regulated frameworks require additional architecture work, documentation, and testing to satisfy.
Data volume & variety
A large or diverse document corpus requires a more sophisticated ingestion pipeline and more thorough retrieval testing to get right.
Number of end users
A system serving 3 people is architected differently from one serving 200. Concurrent load changes hardware requirements, queue management, and deployment complexity.
Get a Real Number
The way to get an accurate quote is to talk.
Tell us about your business and what you're trying to accomplish. We'll follow up with a clear proposal.