Fixed-fee AI agent development.
No monthly seat licenses.
One fixed investment to engineer and deploy your custom AI agent, direct pass-through LLM token billing at 0% markup, and optional month-to-month support. No user tiers, no message credits, and zero proprietary lock-in.
Direct Answer · How xFlow Pricing Works
xFlow AI agents are priced via a one-time project build fee (starting at ₹1,25,000 / ≈ $1,500 USD for a Starter Agent), rather than recurring monthly per-seat subscriptions. Following launch, clients pay actual LLM model token usage directly at cost (typically ₹4,000–₹15,000/month with 0% markup), with optional engineering maintenance available from ₹15,000/month.
Clear, predictable pricing for every automation scope
Choose between a dedicated single-agent build, a multi-agent custom deployment, or ongoing operational maintenance.
Starter Agent
Single-agent automation & speed-to-leadA dedicated, grounded AI agent designed to capture leads in under 45 seconds, automate tier-1 support, or book calendar appointments 24/7.
What's Included
- 1 autonomous agent (Contractor Speed-to-Lead, Support, or Sales)
- Under 45-second response time with emergency SMS routing
- Strict zero-hallucination RAG pipeline with policy guardrails
- Direct calendar booking (Cal.com, Calendly, or CRM webhook)
- Trained on up to 200 pages of service specs and warranties
- Admin dashboard with transcript search and lead exports
- 30-Day Money-Back Satisfaction Guarantee
- Full code, prompt repository, and deployment handover
Custom Build
Multi-agent & enterprise workflowsEngineered for businesses requiring multiple interconnected agents, live database sync, custom tools, or complex integrations.
What's Included
- Everything included in Starter Agent
- Multiple specialized agents with cross-department routing
- Custom API tools with bidirectional database read/writes
- Deep CRM & ERP integration (Salesforce, HubSpot, Shopify Admin)
- Omnichannel reach: Web Chat, WhatsApp Cloud API, and Email
- Custom regression evaluation datasets and accuracy benchmarks
- Dedicated technical architect and tailored enterprise SLAs
Ongoing Support
Post-launch maintenance & tuningOptional managed service to keep your agent accurate, updated with new policies, and continuously tuned from real conversations.
What's Included
- 24/7 uptime monitoring and automated incident response
- Knowledge-base updates as your documentation evolves
- Weekly transcript analysis and answer-quality prompt tuning
- Monthly executive reporting on ticket deflection and ROI
- LLM foundation model upgrades and library security patches
- Dedicated Slack / email channel with engineering team
Note: Brand-new agent builds, new third-party integrations, or custom feature engineering are quoted separately.
Inquire about supportTotal pricing transparency. How costs actually work.
We believe in radical financial clarity. Here is exactly where your budget goes, which costs are fixed, and which costs are variable.
1. One-Time Build Fee
Fixed Investment
A guaranteed, milestone-based price covering complete system architecture, knowledge cleansing, guardrail testing, and deployment. We never bill surprise overtime hours.
- No hourly billing creep
- Fixed deliverables scope
- Milestone payment structure
- Full deployment handover
2. LLM Model Usage
Variable Pass-Through
Direct token consumption from OpenAI, Anthropic, or Gemini passed through at exact cost. Typical monthly spend runs ₹4,000–₹15,000 (≈ $45–$175 USD).
- Billed at exact vendor cost
- Option to link your own API keys
- Upfront volume estimation
- Zero hidden intermediary fees
3. Ongoing Support
Month-to-Month Service
Continuous prompt tuning, knowledge updates, 24/7 uptime monitoring, and model upgrades starting from ₹15,000/month. Cancel or pause anytime.
- No annual lock-in contracts
- Weekly answer accuracy tuning
- Documentation sync as you evolve
- Dedicated engineer communication
4. Third-Party Channels
Direct Provider Billing
If your agent connects to external paid channels like WhatsApp Business Cloud API (Meta conversation fees) or Twilio SMS, those are billed directly by the provider.
- Official Meta Cloud API rates
- Direct carrier telephony pricing
- No xFlow platform surcharges
- Direct provider relationship
Why model usage is billed at true cost with 0% markup
When your agent answers a customer inquiry, it queries a vector database (Pinecone or Qdrant) and generates answers through an LLM (OpenAI GPT-4o, Anthropic Claude 3.5 Sonnet, or Gemini). Unlike subscription software vendors that inflate these costs by 300%–500%, xFlow passes through API tokens at exact cost.
For a business processing 2,000–8,000 monthly customer inquiries, token spend typically totals ₹4,000–₹15,000/month (≈ $45–$175 USD). We calculate an exact forecast for your specific conversation volume during discovery.
Cost Comparison
Traditional SaaS vs. xFlow
A predictable 3-stage delivery roadmap
From our first scoping conversation to production go-live, our procurement process is structured, transparent, and milestone-governed.
Discovery & Technical Scoping
Days 1–3A 20-minute consultation to review your documentation, verify channel compatibility, define escalation logic, and identify required API connections.
Engineering, Grounding & Staging
Weeks 1–3We clean and structure your knowledge base, build deterministic guardrails, connect live tools, and provide a private staging URL for testing.
Production Launch & Full Handover
Go-LiveWe embed the widget onto your live website, configure pass-through token billing, train your staff, and deliver complete prompt repository access.
Engineered for production resilience
We do not wrap simple API calls in a generic template. Every xFlow deployment includes enterprise grounding, evaluation, and logging.
Vector Knowledge Base
Cleansed, chunked, and embedded documentation (up to 200 pages) indexed with hybrid keyword and semantic vector search for rapid retrieval.
Zero-Hallucination Guardrails
Deterministic prompt engineering that confines answers strictly to verified company documentation and executes safe human handoffs.
Live Production Tool Integration
Direct API connection to your CRM, booking calendar (Calendly/Cal.com), helpdesk (Zendesk/Freshdesk), or custom webhooks.
Branded Web Chat Widget
Clean, modern, responsive chat component customized to your exact typography, brand colors, and mobile viewport requirements.
Transcript & Sentiment Analytics
Centralized administrative dashboard to review conversation logs, user sentiment, deflection rates, and unanswered question gaps.
100% IP & Data Ownership
Complete handover of system prompts, guardrail configs, and deployment scripts. No vendor lock-in or proprietary hostage pricing.
Procurement certainty with zero hidden liabilities
Clear commercial terms, transparent billing guarantees, and full intellectual property protection.
Guaranteed Fixed-Price Quotes
What we quote after your discovery call is what you pay. We do not bill unexpected hourly overages or project management surcharges.
Official GST Tax Invoicing
Every Indian contract includes full B2B GST invoices with SAC classification for Input Tax Credit (ITC). Compliant commercial invoices for UAE & global clients.
Zero Per-Seat License Penalties
Traditional chat platforms charge $50–$150 per agent seat per month. With xFlow, your entire internal team can access logs and analytics without paying seat fees.
Direct API Key Transparency
You have the option to plug your own OpenAI or Anthropic API credentials directly, ensuring complete verification of your AI consumption spend.
Pricing questions & commercial terms
Building a production AI agent is an engineering project that tailors knowledge grounding, integrations, and guardrails to your business — not a generic tool you rent per user. Paying once for development gives you 100% intellectual property ownership, zero per-seat licensing penalties, and massive long-term cost savings compared to traditional SaaS seats.
The Starter Agent is a fixed-price engagement live in 2–3 weeks. It includes one autonomous agent (Customer Support, Sales, or Lead Gen), trained on up to 200 pages of verified documentation, one core tool integration (Calendly, CRM webhook, or Helpdesk), a custom branded web chat widget, deterministic guardrails, transcript logging, and complete prompt repository handover.
The underlying AI models (OpenAI, Anthropic Claude, or Google Gemini) and vector database embeddings that power your agent are billed at exact vendor cost with 0% markup. For a typical small-to-midsize business handling several thousand queries monthly, token consumption runs approximately ₹4,000–₹15,000 (≈ $45–$175 USD) per month. You can also link your own direct API keys.
No. Ongoing support is completely optional and billed month-to-month starting at ₹15,000/month. You are never locked into an annual contract. If you have an internal technical team, you can self-host and manage your agent independently following handover.
The 200-page threshold is based on the manual engineering hours required to clean, structure, chunk, and benchmark your knowledge assets. If you have thousands of SKU pages, extensive PDF catalogs, or dynamic documentation, we scope and quote the knowledge ingestion as part of a Custom Build.
Yes. Because xFlow builds modular agent architectures, your system can easily expand. You can add WhatsApp Cloud API, voice channels, deep ERP connectors, or secondary agents at any time. Additional scopes are quoted transparently on a fixed milestone basis.
Yes. All Indian engagements receive formal GST B2B tax invoices with appropriate SAC classification codes, enabling full input tax credit (ITC) for your business. For international clients in the UAE, Singapore, or the US, we provide compliant foreign commercial invoices.
You do, 100%. Upon project completion, we hand over all system prompt repositories, guardrail definitions, integration webhook configs, and conversation databases. xFlow does not lock your data into a proprietary walled garden.
Looking for in-depth comparisons? Project Build vs. SaaS Subscription · Enterprise AI Comparison Guide
Get an accurate, fixed-scope quote
for your AI agent
A 20-minute call is all we need to assess your documentation, determine tool feasibility, and deliver an exact technical scope — not a vague ballpark estimate.
