AI on Your Data
Generic AI hallucinates. AI trained on your own data does not. Three priced paths — RAG, Multi-Agent, or Fine-Tuning — pick the one that matches your problem.
Communication via: Slack, Email, Weekly video calls (Zoom or Google Meet), Loom videos (weekly async demos)
What you get with this package.
ChatGPT does not know your products, your policies, or your processes. So it makes things up — and your team has to correct it every time. Codefree builds custom AI on OpenAI GPT-4 or Anthropic Claude, wired into your documentation via LangChain and a production vector store (Pinecone or Weaviate), that does not have to guess. Three paths depending on the problem: RAG points the model at your PDFs, Confluence pages, SharePoint docs, and web content so every answer is grounded and cited — best for Q&A on existing material. Multi-Agent coordinates specialised agents (retrieval, reasoning, verification) to handle complex, multi-step operational workflows. Fine-Tuning trains a base model on your domain corpus so it speaks your terminology, policy language, and regulatory framing natively — best for legal, medical, and compliance-heavy use cases. AI-native delivery on Claude Code and Cursor compresses the traditional 2 to 3 month RAG or fine-tuning timeline to 2 to 4 weeks without cutting corners on evaluation rigour or safety guardrails. Pick the plan that fits — Codefree will tell you straight on the kickoff call if the wrong one is selected.
Choose your plan.
What you'll receive.
What's included
- Everything in RAG Starter
- Role-specialised agents (retrieval, reasoning, verification, or custom roles)
- Decision support with multiple perspectives combined into a single answer
- Pipeline orchestration and observability dashboard
Not included
- Ongoing hosting and compute costs (AWS, Vercel, GCP — typically 50 to 300 USD per month based on volume)
- Third-party embedding and LLM API fees (OpenAI, Anthropic — typically 30 to 500 USD per month based on query volume)
- Document creation, editing, or content authoring — Codefree works with the knowledge you already have
- Custom customer-facing UI beyond the included admin dashboard (available via the Custom Query UI Build add-on)
- Data migration from legacy systems (SharePoint, Confluence, legacy CMS — priced separately as a discovery engagement)
Optional add-ons.
Extend your package with these optional extras to get even more value.
Integrations
Additional Data Source Integration
PopularIngest documents from one more source system beyond what is included — Confluence, SharePoint, Google Drive, Notion, or a custom REST API.
UI
Custom Query UI Build
Customer-facing or internal query UI beyond the included admin dashboard — Slack bot, Teams bot, or a branded web widget with your styling.
Support
Extended Support Retainer
90 additional days of bug-fix and tuning support beyond the included 30, plus one review and adjustment cycle per month based on production performance data.
Add-ons can be selected during booking or added later. Start booking →
Why choose this package.
Key capabilities.
Is this the right fit for you?
This package is designed for specific types of people facing specific challenges. See if your situation matches.
Knowledge Manager
Your team's knowledge is scattered across hundreds of documents nobody can find.
Your Challenges
- •Employees waste hours searching for information across Confluence, SharePoint, Notion, and Google Drive
- •Critical knowledge is locked in documents that nobody reads or knows exists
- •Same questions get asked repeatedly because answers are not accessible or discoverable
- •Onboarding new hires takes weeks because tribal knowledge lives in senior team members' heads
What You Want
- •Make institutional knowledge instantly accessible to everyone via a single query interface
- •Reduce time spent searching and recreating known solutions across the organisation
- •Turn static documents into an intelligent, queryable resource with source citations
Deploy an AI that answers questions from your actual documents — with citations, so people trust and verify the answers, and your team stops answering the same question 50 times.
Need Something Different?
Open-ended creative tasks (use generic AI instead) or use cases where you do not yet have documented knowledge to train on. Codefree makes AI use your knowledge — we do not create the knowledge for you. Also not a fit for simple support chatbots without a knowledge grounding requirement — use the AI Agents & Chatbots package for that.
Perfect for these scenarios.
Technologies we use.
What you'll need to provide.
- 1.Existing documentation (PDFs, Word docs, Confluence, SharePoint, or web pages) that captures the knowledge you want the AI to use
- 2.Clear use case definition — the specific question types the AI needs to answer at day one
- 3.Admin access to your document sources (SharePoint, Confluence, Google Drive, S3, or equivalent) with read scope
- 4.Defined user access and permission requirements for the query interface (public, internal-only, role-based)
How we work together.
Discovery, document audit, and architecture planning
Codefree interviews 2 to 3 people on your team, audits a sample of your documents (PDF, Confluence, SharePoint), and produces a written architecture plan you sign off on — including model choice (OpenAI, Anthropic), vector store choice (Pinecone, Weaviate), and evaluation criteria.
Document ingestion and vector store indexing
Codefree processes your documents through the ingestion pipeline (format-aware parsing, chunking, embedding) and indexes them into the vector store. End of Week 1 you can query the raw index in a staging environment.
AI-native RAG pipeline development and retrieval tuning
Build the retrieval pipeline on Claude Code and Cursor, tune chunking and search parameters against your specific query patterns, wire in confidence scoring and human-in-loop checkpoints. End of Week 2 you get a working prototype to test with real queries.
Quality testing, edge case handling, and production deployment
Adversarial testing on out-of-scope queries, hallucination edge cases, and low-confidence handoff. Deploy to your live environment with a rollback plan. 30-day post-launch bug-fix support kicks in from go-live.
What our clients say.
“Scholar Freedom was designed to meet two missions -- open credible research to the public and ensure academics are properly paid. Codefree.io were engaged as a no-code agency to build the platform. They began to bring the dream to life with their eye for detail, imagination, and knowledge.”
“We worked closely with Codefree developers for over a year, to develop complex apps, tools & solutions for our global client's requirements. Codefree understands the data flows, security & other technical aspects for a multi-language, multi-country, multi-platform environment.”
“CodeFree.io has been an absolute game-changer for our web app project. Their expertise in no-code development is unmatched, and their dedication to our success has been outstanding.”
Common questions.
What document formats do you support for RAG ingestion?+
How do you handle document updates in the RAG system?+
Can users see where the AI answers come from?+
How accurate is the system in production?
Ready to get started?
Book this package now and let's bring your project to life. Clear pricing, fixed scope, no surprises.
Not quite right? Explore bespoke solutions for custom requirements.
You might also like.
AI Agents & Chatbots
Custom AI agents and chatbots built on your data and integrated with your existing SaaS tools. Cut manual workload by 60 to 80 percent. Deployed in 2 to 4 weeks with AI-native velocity — send us your brief today.
AI Voice & Call Center
Three priced paths to voice AI — Inbound voice agent, Outbound campaigns, or Voice inside your app. CRM-integrated, multi-accent, with human escalation built in. Delivered in 4 to 8 weeks with AI-native velocity.