QUASIHUMAN / v0.4
Platform

One agent across every system you run, with a memory that outlasts the conversation.

What follows is what a QuasiHuman operator is actually made of: tools that span your whole landscape, interfaces it builds for you mid-turn, three kinds of memory that survive the week, every doorway it arrives through, and the runtime underneath all of it.

01 / FEDERATIONOne colleague, every system

A single-system bot is a commodity. Reasoning across your whole landscape is the product.

The agent doesn't hand you between four bots. It reads the account, files the claim, checks the finance position and evaluates the policy inside a single conversation — carrying what it learned between each step.

Every system reaches it as curated, task-level tools: typed, classified by blast radius, every mutation gated. Not raw database access. Not screen-scraping.

CAPABILITY BOUNDARY · YOUR SYSTEMS, YOUR RULES
CRMERPFinanceTicketingWarehouseSchedulingData warehouse + bespoke
SINGLE TURN · TRACE
SAMPLE
USER Register the warranty claim for A-4192, then check the customer's finance standing before approving a loaner.
crm.account.lookup A-4192 94ms ok
warranty.claim.create WC-88213 310ms ok
finance.standing.check A-4192 140ms 2 flags
policy.evaluate loaner_approval 12ms HELD
◆ APPROVAL REQUIRED · BLAST RADIUS HIGH
Loaner approval exceeds the autonomous limit for this role. Routed to @supervisor — agent parked, full state preserved, resumes on decision.
02 / INTERFACEGenerated interfaces, not generated code

It builds the interface, not just the answer.

When a paragraph won't do, the agent builds you an interface — a form, an approval, a triage queue — right inside the conversation. Ask for it once and it's saved: you reopen it months later on live data, with no model call anywhere on the path.

THE INVARIANT

A surface is data, never code. The model authors a validated declarative tree against a closed component registry — no HTML, no iframe, nothing generated ever executes — and every capability it reaches is re-authorised against your policy on every call.

AGENT-AUTHORED · MID-TURN AWAITING YOU
Approve loaner · account A-4192
Vehicle class Compact
Duration 3 days
Cost centre FLEET-EU-02
Approve Decline Ask for more
turn parked · resumes on your choice · your answer is re-spotlit as untrusted input
SAVED SURFACES
Morning triageopened 41× Exceptionsopened 18× Fleet SLAopened 7×
live data · no LLM call on the open path
03 / BRAINIt compounds

A worker is not a transcript. It's something that survives the week — and teaches the next one.

DURABLE EXECUTION

Kill the server mid-task. It resumes where it stopped.

Every conversation is a durable workflow, checkpointed at every step — and it never repeats the side effects it already committed. Work that parks for a human decision on Friday picks up on Monday holding everything it knew.

14:02:41 tool.exec finance.post ok
14:03:02 worker lost node terminated
14:03:07 resumed step 7 of 12
14:04:15 tool.exec report.compile ok
MEMORY & SKILLS

It remembers Tuesday.

Three kinds of memory — what happened, what's true, and how work actually gets done here — retrieved by meaning, keywords, entities and recency across four tiers. A procedure that works is kept as a skill, and every time it's used the outcome is recorded.

EPISODIC A-4192 escalated over a delayed part on 14 Aug.
SEMANTIC Fleet accounts approve loaners up to £400 without finance sign-off.
PROCEDURAL Fleet warranty claims skip the pricing check.
ORGANISATION BRAIN

What one person learns, the company keeps.

Agents maintain a knowledge base for your organisation out of the work itself — integration guides, process docs, why a decision went the way it did. Every page and every memory is authorised at query time against your hierarchy: level, team, role. Someone leaves; what they taught the system stays.

ORGANISATION every team, every level
└─ FINANCE team-scoped
└─ CONTROLLERS level-gated
└─ YOU personal memory
04 / SURFACESWhere your agents show up

One agent. Every doorway. The runtime is the product — the chat window is just one way in.

SHIPPED
Embedded widget

One script tag drops an agent into your own applications, authenticated against your service — not a third-party bubble in the corner.

SHIPPED
System events

Any upstream service, webhook or cron can start an agent run programmatically. No human has to be in the room for work to happen.

SHIPPED
Co-browse

The agent perceives the page your user is actually on and acts on it with them — surviving navigation, asking before it touches anything.

IN BUILD
Live surfaces

When a paragraph won't do, the agent builds a small interactive interface in the conversation — and you can keep it.

# IN BUILD
Slack

Direct messages and channel mentions, streaming its answer into the thread where the work is already being discussed.

IN BUILD
Voice

Embedded voice and inbound telephony on the same runtime — the caller talks to the agent that already knows the account.

IN BUILD
Routines

Recurring agentic work on a schedule. The Monday reconciliation runs itself and only escalates when something looks wrong.

ROADMAP
Email

Agents with a mailbox: reading, searching, replying — and inbound mail that routes straight to the agent expecting it.

Same memory, same tools, same supervision, same audit trail — whichever door the work comes through.
05 / RUNTIMEThe parts underneath

Not six chat windows. One runtime, with the parts a real operator needs.

QH-01
Sub-agents

It delegates, parks, and picks the work back up.

Big jobs fan out to specialists that run independently and report back. The parent waits without burning budget, then continues with everything the children found.

agent.root · warranty escalation ├─ sub.research parts availability done ├─ sub.finance standing check done └─ sub.drafting customer comms running
QH-02
Objectives

Goals that outlive the conversation.

A multi-step objective is a durable plan, not a prompt. It survives across days and chats, tracks its own dependencies, and reports progress against the outcome you asked for.

QH-03
Artifacts

Reviewable output, not chat scrollback.

Work lands as versioned artifacts your team can diff, comment on, approve or roll back. The record of what an agent produced is a document, not a transcript you have to scroll.

QH-04
Code sandbox

It writes code, then runs it somewhere safe.

When the answer needs computation rather than a plausible-sounding paragraph, the agent writes code and runs it in an isolated microVM — no network trust, no access to your systems except through the tools you granted.

Bring us the workflow that nobody wants to own.

The platform earns its keep against a messy landscape, not a demo. Tell us which systems have to agree with each other and we will map the capability boundary with you.