The Core of CorpusAI
Every business has knowledge that exists nowhere else — contracts, policies, case files, internal procedures. A corpus turns that knowledge into an AI that understands your world.
Generic AI knows the internet. Your corpus makes AI know your business.
When senior employees leave, their knowledge walks out the door. A corpus captures decades of institutional knowledge — decisions, precedents, rationale — and makes it instantly searchable for every team member.
Your competitors can use the same AI models you do. They can't replicate your data. An internal corpus gives you answers grounded in your contracts, your case history, your internal policies — knowledge no competitor has access to.
Ask ChatGPT the same question twice and you get different answers. Ask your corpus and every answer is grounded in specific documents with citations. Your team gets consistent, verifiable information every time.
Knowledge workers spend 20% of their time searching for information. A corpus gives instant answers from your document base — no more digging through folders, email threads, or asking colleagues.
New hires take months to learn your systems, processes, and history. With a corpus, they can ask questions in natural language and get answers from your full document base on day one.
Every document you add makes your corpus smarter. Unlike static databases, a corpus continuously improves — new precedents, updated policies, fresh case files all become part of your AI's corpus.
Knowledge That Stays
When a senior partner retires or an experienced engineer leaves, decades of institutional knowledge walk out the door. A corpus captures that expertise — decisions, precedents, rationale — and makes it searchable for every team member.
Every document you add makes your corpus smarter. Unlike static databases, this knowledge grows and compounds over time.
Choose the strategy that fits your business. Start simple, scale up as you grow.
Upload your own documents. Build a completely private CorpusAI corpus that only your team can access. Perfect for companies with proprietary content that shouldn't be mixed with external sources.
Ideal for
Example query
“What is our policy on remote work for employees in the Nordic region?”
Answers from: Employee Handbook v4.2, Nordic Office Policy 2025
Combine your private documents with expert-curated marketplace corpora. Your internal knowledge meets industry-grade reference material — all queryable in a single question.
Ideal for
Example query
“Does our client contract comply with barneloven § 48 regarding the child’s best interest?”
Your files: Client Contract #4821 · Marketplace: barneloven § 48, HR-2020-1843-A
Don't have documents to upload yet? Start with expert-curated corpora from the marketplace. Get instant access to industry knowledge — legislation, regulations, case law — without uploading a single file.
Ideal for
Example query
“What are the key obligations under the EU AI Act for high-risk systems?”
Answers from: EU AI Act Corpus — Articles 6-49, Annex III
Three paths to get your documents indexed. Pick one or combine all three — your AI gets smarter with every document.
Upload PDFs, Word documents, or text files — one or fifty at a time. Drag them in, add metadata, and watch them get chunked, embedded, and indexed in real time. No technical skills required.
PDF, DOCX, TXT, HTML, MD — up to 50 MB each
Point a scraper at any public data source. Configure it to run daily or weekly, and your corpus stays current automatically. Every new regulation, guideline, or publication gets indexed without lifting a finger.
SSRF-protected — configurable depth & page limits
Integrate with your existing systems. Push documents programmatically, tag them with rich metadata, and let the pipeline handle the rest. Perfect for CMS integrations, document management workflows, and CI/CD pipelines.
X-API-Key auth — plan-gated document limits
Every document can carry metadata — author, tags, source URL, publication date, category. This isn't just filing. It's how your AI learns to filter, prioritize, and cite the right source for the right question.
Amplified Knowledge
Your internal documents contain the answers to "what did we do?" Marketplace corpora contain the answers to "what should we do?" Combined, they let you ask questions that span both worlds.
Check your contracts against current legislation. Verify your policies meet GDPR requirements. Cross-reference your case files with Supreme Court precedent — all in a single query.
See how adding marketplace corpora amplifies the value of your internal documents.
Answer from your data
Grounded in your contracts, policies, case files
Full privacy
Zero external data mixing
Exact document citations
Page numbers, paragraphs, file names
No regulatory context
Your data doesn't include the laws it's subject to
No cross-referencing
Can't check compliance against external standards
Everything above, plus…
All internal corpus benefits included
Regulatory cross-referencing
Check your docs against legislation automatically
Compliance gap analysis
"Does our privacy policy cover all GDPR Article 13 requirements?"
Multi-source citations
Answers cite both your docs and relevant regulations
Always up-to-date regulations
Marketplace corpora maintained by domain experts
The ultimate competitive advantage: a fine-tuned AI model trained specifically on your corpus. Not just retrieving your documents — thinking like your organisation.
Upload documents, configure scrapers, grow your corpus
We fine-tune an AI model on your corpus using GPU infrastructure
Test against your domain questions, iterate until quality matches expectations
Your model goes live — faster, cheaper, and smarter than generic AI
3× faster responses
A fine-tuned model doesn't need to search your corpus at query time — the knowledge is baked into the model weights.
Lower per-query cost
No vector search overhead per query. A smaller, specialised model outperforms a larger generic one on your domain.
Domain vocabulary mastery
The model learns your industry's jargon, abbreviations, and terminology. It understands "BL § 48" means barneloven, not a product code.
Your model, your control
The fine-tuned model runs on EU infrastructure. You own the output. It can be deployed privately or integrated via API into your existing systems.
Fine-tuning and RAG (Retrieval-Augmented Generation) aren't competing approaches — they're complementary. Use both for the best results.
Searches your corpus at query time. Always up-to-date. Great for fact-finding, document lookup, and citation-heavy queries. Works immediately after upload.
Knows your domain deeply. Faster responses. Better at reasoning, analysis, and synthesis. Requires periodic retraining when your corpus changes significantly.
The fine-tuned model understands your domain vocabulary and reasoning patterns, while RAG ensures it always cites the latest documents. Deep understanding + real-time accuracy.
Already using CorpusAI? Your existing corpus is automatically the training data for fine-tuning. No extra preparation needed — just contact us when you're ready to take the next step.
See how different businesses use corpus strategies to solve real problems.
Law Firm
Handles custody, divorce, and child welfare cases across Norway.
Corpus Setup
Result
Lawyers ask one question and get answers citing both their own case history and relevant legislation. New associates become productive in days, not months.
SaaS Company
Expanding to EU market, needs to ensure AI Act and GDPR compliance.
Corpus Setup
Result
Compliance team can instantly check if product features meet EU AI Act obligations. Gap analysis that took weeks now takes minutes.
Enterprise + Fine-Tuning
Processes thousands of claims and needs fast, consistent assessments.
Corpus Setup
Result
A dedicated AI model trained on their claim history. Assessors get instant, consistent recommendations. Processing time reduced by 60%. The model speaks their exact terminology.
Start with a free corpus. Grow with marketplace corpora. Graduate to a fine-tuned model. Your data, your pace.