DODIL CLOUD

Turn your data into
knowledge.

Connect your sources. Transform, retrieve, and act on your data — all from infrastructure that never leaves your environment.

Sovereign by design. AI-ready in minutes.

Get Early Access Explore Products

Step 01

Connect your data

Create an S3-compatible knowledge bucket and point it at your sources. Every bucket becomes a complete knowledge instance — storage, pipelines, vector index, and API, all in one.

Step 02

Transform & index

Documents are chunked, summarized, embedded, and indexed through massive parallel pipelines. Delta-sync re-processes only what changed. Your data becomes AI-ready without leaving your environment.

Step 03

Query, retrieve, act

Hybrid semantic search, RAG pipelines, and agentic workflows — all through one API. Ask questions in natural language, feed context to your models, and build AI that acts on your knowledge.

IngestAny file, any source

TransformChunk, summarize, embed

RetrieveSemantic + hybrid search

DeployRAG & agents in minutes

Quick startProvision a bucket via the dashboard and start ingesting in minutes. No code required.Try the dashboard Use the SDKPython and Node.js SDKs for programmatic ingestion, search, and pipeline control.View docs Build workflowsDefine multi-step AI pipelines with Scriptum — compile-time validated, resumable, streaming.Explore Scriptum

Everything you need to build
AI-powered knowledge systems

Four products. One platform. From storage to search, orchestration to compute.

K3 · Knowledge Storage

Augmented S3 that understands your data

Every bucket is a knowledge instance. Upload files, define ingest rules, and get hybrid semantic search — vectors + BM25 — out of the box.

Explore K3

Ignite · Compute

The AI self-tooling platform

Serverless functions that agents create, test, and call as MCP tools. ~200ms warm cold starts, Python and Rust runtimes, massively parallel by default.

Explore Ignite

Scriptum · Orchestration

LLM workflows that actually compile

A purpose-built DSL with 15 primitives. Full type checking before execution, resumable threads, streaming pipes, and first-class human-in-the-loop.

Explore Scriptum

VBase · Vector DB

Vector search, served

A managed vector database for RAG, semantic search, and agentic memory. Serverless pools, automated backups, multi-tenant isolation, and a drop-in pymilvus client.

Explore VBase

Built for industries where
data can't leave.

Sovereign AI knowledge infrastructure for regulated enterprises — and the tech companies building for them.

M&A Due Diligence

Find Any Clause Across 10,000 Contracts

Legal teams spend months hand-reviewing contracts during due diligence. Pre-announcement M&A data cannot leave the firm's jurisdiction.

Time saved

6 months of manual reviewa single weekend

Hybrid search catches semantic clause variants that keyword review misses entirely.

Contract extraction pulls parties, dates, governing law, and renewal terms from every PDF — overnight
OCR + Layout makes scanned 1990s contracts as searchable as today’s
Hybrid search returns semantic clause variants keyword review misses

Contract extractionOCR + LayoutEntities + PIISummarizationClaim/Fact extraction

Data room stays in your jurisdiction — Magic Circle and PE-grade.

Get Early Access

Live pipeline

Uploading 10,000 contracts to K3 bucket

Contract extraction + OCR + entity tagging

Search: "change of control provisions with carve-outs"

47 matches across 12 contracts

Sovereign infrastructure
for AI workloads.

Your data plane runs in your datacenter. Our control plane is in London — expanding to Middle East and Asia.

LondonActive

Middle EastPlanned

AsiaPlanned

Get Early Access

DODILFrom data to intelligence.

Home Pricing Progress

Turn your data into knowledge.