01
Chat and writing
Answer questions, draft text, explain documents, summarize.
§ PLATFORM OVERVIEW
One platform.
Every kind
of AI.
Writ gives your developers one simple way to use eleven different kinds of AI. One login. One set of rules. One set of logs. And every kind of AI uses the same approach, so learning one means learning them all.
§ CAPABILITIES
Eleven things Writ can do.
Each capability is available through the same simple web address. Your developers call one, call another, or chain them together — and the experience feels identical. The goal is to let your team focus on the mission, not on stitching products together.
02
Prediction
Sort records into categories; forecast outcomes from features.
03
Image understanding
Find objects, recognize scenes, describe or tag photos and video.
04
Search with AI
Find the right document, pull the right paragraph, cite the source.
05
Graph-aware search
Follow relationships between people, places, events, and records.
06
Speech
Transcribe audio. Read text aloud. Separate speakers.
07
Mixed inputs
Understand images and words together — document photos, charts, scans.
08
Time-series
Forecast a trend. Flag the reading that doesn't belong.
09
Ranking
Sort a long list so the important items come first.
10
Learning from feedback
Improve an assistant or model with human preference data.
11
Simulation
Run what-if models for training and scenario planning.
12
Agents
Plan a multi-step task, choose the right tool, and execute it.
§ HOW IT WORKS
Five steps. Every time.
Whether a developer is asking a question, running a prediction, searching documents, or coordinating an agent — the steps are the same. That's the point of having one platform.
§ HOW A REQUEST FLOWS
STEP · 01
Request
A developer or an AI tool asks Writ to do something.
STEP · 02
Verified
Writ checks who's asking and what they're allowed to do.
STEP · 03
Routed
The request goes to the right kind of AI — chat, prediction, search, and so on.
STEP · 04
Answered
The AI produces a result.
STEP · 05
Signed & logged
The answer is signed. A complete audit record is filed.
STEP · 01
Request
A developer or an AI tool asks Writ to do something.
STEP · 02
Verified
Writ checks who's asking and what they're allowed to do.
STEP · 03
Routed
The request goes to the right kind of AI — chat, prediction, search, and so on.
STEP · 04
Answered
The AI produces a result.
STEP · 05
Signed & logged
The answer is signed. A complete audit record is filed.
STEP · 01
Request
A developer or an AI tool asks Writ to do something.
STEP · 02
Verified
Writ checks who's asking and what they're allowed to do.
STEP · 03
Routed
The request goes to the right kind of AI — chat, prediction, search, and so on.
STEP · 04
Answered
The AI produces a result.
STEP · 05
Signed & logged
The answer is signed. A complete audit record is filed.
ENCRYPTED END-TO-END
§ PLATFORM MAP
Every service. Every connection.
This is the interactive map of the platform — every service and every relationship between them. Hover a box to see what that piece does. Hover a line to see how two services talk to each other. Click to pin your view in place.
§ PLATFORM MAP — hover a box or a line
▸ View as list
CLIENT
App / SDK / CLI
Applications you build. Your mobile app, your web app, your Python or TypeScript service. They import the Writ toolkit and call the platform over the network.
CLIENT
AI Coding Tool
Claude Desktop, Cursor, Windsurf, and other tools that speak the Model Context Protocol. A developer runs one command and their editor can use every Writ capability.
EDGE
Edge Gateway
The front door. Terminates encrypted traffic, enforces rate limits, and passes validated requests to the control plane. Speaks hybrid future-proof encryption (X25519 + ML-KEM-768).
CONTROL PLANE
API Gateway
The single OpenAPI 3.1 surface at /v1/*. Every capability — chat, prediction, search, speech, agents — lives here. One auth model, one error model, one audit schema, one rate-limit policy.
CONTROL PLANE
MCP Director
The tool plane. Exposes every Writ capability as a Model Context Protocol tool. One endpoint; namespaced tools; OAuth 2.1 scopes; everything audited.
IDENTITY
Keycloak
OIDC identity broker. Handles CAC / PIV / enterprise SSO, federation to external IdPs, and token issuance.
IDENTITY
SPIRE
Workload identity. Every service in the cluster gets a SPIFFE identity issued by SPIRE; internal mTLS is verified against it.
IDENTITY
OpenBao
Secrets and key management. HSM-backed. Replaces HashiCorp Vault under a permissive license.
POLICY
OPA
Policy decisions. For every request, OPA evaluates tenant, classification, purpose, and release markings against the rules your admins wrote.
AI RUNTIME
vLLM
Generative model server. Runs large language models with efficient batching, streaming, and multi-tenant isolation.
AI RUNTIME
KServe
Inference server for predictive, classical-ML, and computer-vision models. Scales to zero when idle; spins up when work arrives.
AI RUNTIME
Triton
High-throughput inference for vision and multimodal models, with TensorRT-LLM backends for NVIDIA GPUs.
AI RUNTIME
LangGraph
Agent runtime. Implements plan / act / reflect loops, tool calling, and multi-step workflows with persistent checkpoints.
DATA PLANE
PostgreSQL
The main database. Stores conversations, agent checkpoints, user records, and vector embeddings for AI search (via the pgvector extension).
DATA PLANE
OpenSearch
Full-text and hybrid search. Documents with classification tags; policy-enforced retrieval.
DATA PLANE
MinIO
Object storage. Model weights, training data, evidence bundles. Works offline, synchronizes when a link opens.
DATA PLANE
NATS JetStream
Messaging and event streaming. Feeds audit records, model-request queues, and cross-service events.
SUPPLY CHAIN
Sigstore
Signing and transparency. Every container image, every release, every audit bundle is signed with post-quantum-safe keys and logged to a public record.
OBSERVABILITY
Observability
Metrics, logs, and traces — the operations view of the platform. Dashboards your team already knows how to read. Every service reports; nothing is black-boxed.
§ WORKS WITH YOUR TOOLS
Plays well with modern AI assistants.
Developers who already use assistants like Claude Desktop, Cursor, or other modern coding tools can connect them to Writ with a single command. The same rules and security that protect the platform protect those connections.
Every connection is authenticated against your identity system. Every request is logged. Every answer is auditable.
// Connect your AI tool
writ
mTLS · hybrid-X25519+ML-KEM-768
$ writ connect claude-desktop § WHERE IT RUNS
One install. Three sizes.
Writ is the same software from a developer's laptop all the way to a full government data center. A program typically starts small and grows as the mission matures — with no rewrite along the way.
SIZE · 01
writ.profile.dev
Laptop
A single file you run on a developer's computer. Works offline. No special hardware needed. Good for a first demo.
- ◆Any laptop, 8 GB of memory
- ◆Works without an internet connection
- ◆Seeded with sample data
SIZE · 02
writ.profile.edge
Single server
One hardened server in the field. Runs on standard hardware — most brands of Intel, AMD, and ARM. Supports limited AI chips.
- ◆8–16 CPU cores
- ◆Optional AI accelerator
- ◆Ready for air-gapped networks
SIZE · 03
writ.profile.cluster
Full data center
A full cluster in a data center. Built for high-sensitivity government workloads. Supports multiple tenants and cross-site operation.
- ◆Kubernetes cluster, standard or hardened
- ◆Multiple AI accelerators
- ◆Security-hardened base images only
§ FOR TECHNICAL READERS