Cloud,
rewritten for AI.

Compute, models, memory, and workflows — auto-wired on a sovereign UK/EU cloud. One auth, one bill, live on day one.

❯dodil up
resolving sovereign regionlon1 · uk
ignitecompute12ms
modelscatalog08ms
k3 · objectsstorage15ms
k3 · tableshtap21ms
k3 · vectorsearch19ms
scriptumworkflows09ms
cloud live
❯

Built on Dodil:

RAG

a RAG app over your own PDFs.

Drop docs in K3 — Scriptum chunks and embeds them, Vector serves hybrid search.

WAREHOUSE

a scalable analytical warehouse.

K3 Tables is serverless HTAP. Open it in a spreadsheet, query it in SQL, scale to terabytes.

VECTOR DB

a managed vector DB that scales to zero.

Multi-tenant indexes, hybrid search, durable backups. Pay for queries, not idle infrastructure.

AGENT

agents that survive crashes and approvals.

Typed pipelines, durable state, and pause/resume for humans in the loop.

one object queried three ways — no ETL.

K3 stores it once. Vector reads it semantically, Tables runs SQL on it, Objects serves the bytes.

DEPLOY

a function or container with one command.

Python, Node, Dockerfile, Buildpacks, GitHub, or BYO image. Scales to zero, billed by the millisecond.

MODELS

Chat, Embeddings, and Rerank from one API.

OpenAI-compatible drop-in. Point your existing SDK at Dodil — keep your code, sovereign data path.

EVENTS

every upload as a live pipeline.

S3-compatible PUT fires Scriptum on the way in — extract, embed, route, file, notify.

RAG

a RAG app over your own PDFs.

Drop docs in K3 — Scriptum chunks and embeds them, Vector serves hybrid search.

WAREHOUSE

a scalable analytical warehouse.

K3 Tables is serverless HTAP. Open it in a spreadsheet, query it in SQL, scale to terabytes.

VECTOR DB

a managed vector DB that scales to zero.

Multi-tenant indexes, hybrid search, durable backups. Pay for queries, not idle infrastructure.

AGENT

agents that survive crashes and approvals.

Typed pipelines, durable state, and pause/resume for humans in the loop.

one object queried three ways — no ETL.

K3 stores it once. Vector reads it semantically, Tables runs SQL on it, Objects serves the bytes.

DEPLOY

a function or container with one command.

Python, Node, Dockerfile, Buildpacks, GitHub, or BYO image. Scales to zero, billed by the millisecond.

MODELS

Chat, Embeddings, and Rerank from one API.

OpenAI-compatible drop-in. Point your existing SDK at Dodil — keep your code, sovereign data path.

EVENTS

every upload as a live pipeline.

S3-compatible PUT fires Scriptum on the way in — extract, embed, route, file, notify.

Six primitives.
The glue ships with them.

Three families, one connected stack. Compute runs your code and ours. K3 is your memory — objects, tables, and vectors as one bucket. Scriptum wires it all together.

Compute

Ignite

Run code, anywhere.

Serverless functions and containers. Bring whatever you've got — we build, deploy, and scale it.

DockerfileBuildpacksGitHubBYO Image

$ ignite deploy ./fn

Explore Ignite

Ignite Models

Call a model, not a vendor.

Pre-built model catalog. OpenAI Chat & Embeddings, Cohere Rerank, custom Predict — one API, OpenAI-compatible drop-in.

POST /v1/embeddings

Browse catalog

Memory

K3 Objects

S3 that understands your data.

Wire-compatible with S3 — swap your endpoint, keep your SDK. Every upload can fire a Scriptum pipeline automatically.

$ aws s3 cp ./pdfs s3://my-bucket

Explore K3 Objects

K3 Tables

Serverless SQL → HTAP.

Serverless HTAP, built in Rust. Open in a spreadsheet, query in SQL, or write through pipelines.

SELECT * FROM bucket.docs

Explore K3 Tables

K3 Vector

Hybrid vector search.

Semantic + keyword search with rank fusion, attached straight to your K3 Objects bucket. Built for RAG and search.

vector.search('contract clause')

Explore K3 Vector

Workflows

Scriptum

Typed pipelines. The glue that wires every primitive together.

A purpose-built DSL with 15 primitives. Compile-time validation, resumable threads, streaming pipes. Triggered by uploads, schedules, or events — writes to Tables and Vector automatically.

PDFk3.objects.on

Chunkchunk.text

Embedembed

Vectork3.vector.upsert

Explore Scriptum

Pricing

Pay for what runs.

Start free — no card required. Every account opens on the Free plan with hard caps on every meter, so there are no surprise bills. Upgrade to Early Bird ($19.99 / month, first 100 only) for 5–17× more headroom. The per-meter rates below preview the PAYG model coming next.

See Free, Early Bird & PAYG preview →

Scales to zero

Pay for bursts, not for idle.

Ignite Apps and K3 Tables compute autoscale pods down to zero between requests. Catalog Models, vector reads, storage — everything else is metered per token, per GB, per request.

VM baseline

You payIdle — free

IgniteCatalog Models

Per tokenhosted LLMs · × per-model multiplier

Per CPU-secself-hosted models · same as Apps compute

Kimi · embeddings · rerankers · OCR · transcription · vision — every model has its own rate. See the rate sheet for the full list.

IgniteApps

$0.0432/ vCPU-hour

$0.0060/ GB-RAM-hour

Serverless containers. Pods autoscale to zero between requests.

K3 TablesQuery compute

$0.0432/ vCPU-hour

$0.0060/ GB-RAM-hour

Same engine as Apps. You pay only when a query is running.

Free is the default plan. Early Bird ($19.99 / mo) is the paid upgrade while spots last; PAYG opens after the first 100. Need more? Talk to us.

See pricing

Sovereign by design

Your data,
in your region.

Compute, storage, indexes, and pipelines run in the region you pick — and stay there. Live in the UK and EU, with the Middle East and Africa next.

SOC 2In progress

ISO 27001In progress

GDPR-ready

Data residency enforced

Start free

One stack instead of five.

The same AI app on a hyperscaler means assembling six services and writing the glue. On Dodil, the glue is already there.

On a hyperscaler

S3Object storage

Athena / RedshiftSQL warehouse

OpenSearch / PineconeVector search

Lambda / FargateFunctions / containers

Step Functions / GluePipelines

Bedrock / SageMakerModel inference

glue YAML

On Dodil

Dodil stack

K3 Objects

Object storage

K3 Tables

SQL warehouse

K3 Vector

Vector search

Ignite

Functions / containers

Scriptum

Pipelines

Ignite Models

Model inference

One authOne billWired on day one

Six services, six bills, and weeks of YAML to wire them together — versus one stack, one auth, wired on day one.

For developers

Start building
in minutes.

SDK, CLI, and a free tier. Ship a function, query a vector, run a pipeline — all without leaving your editor.

Start free Talk to usNo card required · Cancel anytime

~/my-app

$ curl -fsSL cli.dodil.io | bash
$ dodil ignite deploy ./fn
✓ function live
$ dodil k3 vector search
✓ 12 hits · 28ms
$

DODIL

Sovereign AI cloud. Six primitives, one bill, wired on day one.

Compute & AI

IgniteServerless functions & containers

Ignite ModelsOpenAI-compatible model catalog

Storage

K3 ObjectsS3-compatible object storage

K3 TablesDelta Lake HTAP SQL

K3 VectorHybrid dense + sparse search

Workflows

ScriptumTyped pipeline DSL

Resources

CLIInstall the dodil command line

Talk to usSales · custom plans · sovereign deployments

Regions

UKLiveEULiveMiddle EastSoonAfricaSoon

Compliance

SOC 2In progressISO 27001In progressGDPR-readyData residencyEnforced

Privacy Terms Security Acceptable Use DPA SubprocessorsBuilt for the AI era.

Cloud,rewritten for AI.

Built on Dodil:

a RAG app over your own PDFs.

a scalable analytical warehouse.

a managed vector DB that scales to zero.

agents that survive crashes and approvals.

one object queried three ways — no ETL.

a function or container with one command.

Chat, Embeddings, and Rerank from one API.

every upload as a live pipeline.

a RAG app over your own PDFs.

a scalable analytical warehouse.

a managed vector DB that scales to zero.

agents that survive crashes and approvals.

one object queried three ways — no ETL.

a function or container with one command.

Chat, Embeddings, and Rerank from one API.

every upload as a live pipeline.

Six primitives. The glue ships with them.

Run code, anywhere.

Call a model, not a vendor.

S3 that understands your data.

Serverless SQL → HTAP.

Hybrid vector search.

Typed pipelines. The glue that wires every primitive together.

Pay for what runs.

Pay for bursts, not for idle.

Your data, in your region.

One stack instead of five.

Start buildingin minutes.

Cloud,
rewritten for AI.

Six primitives.
The glue ships with them.

Your data,
in your region.

Start building
in minutes.