Ablo is collaboration infrastructure for AI agents

For teams building products where agents, apps, services, and people write to the same rows: one API coordinates every writer, while your data stays in your own Postgres.

Run npm install @abloatai/ablo, then npx ablo init, and follow the quickstart.

import { ablo } from "@/ablo/client";

await using held = await ablo.records.claim({ id });

const summary = await summarize(held.data);

await ablo.records.update({

id, data: { summary }

});

One typed client claims a row, does the slow work, and writes it back. Runs wherever your code runs, with any framework or none.

Context tells an agent what it knows.

Shared state tells it what is currently true.

Stale-read detection

You can’t overwrite a change you never saw. Edits built on old data are caught, not committed.

Authorship

Nothing changes anonymously. Every edit carries who made it and what they based it on.

Claims

One editor at a time. Everyone else waits their turn instead of overwriting each other.

Real-time sync

Everyone sees the same state the moment it changes, with no refreshing or stale copies.

Expose the same live state to every agent working together.

checkout-service>Bundle discounts

M

A

01

02

03

04

Commit

Up to date4

Bundle discounts12m

Ready to write the discount lines now.01M

Cover total paths31m

That one asserts on a number src/pricing.ts is still changing, so I will take the file when it settles.02A

Tighten the schema1h

Carts and lines are declared per field now, so the new models read beside them.03M04

Audit rounding3h

Holding off on the change. I am reading src/schema.ts to pick up the shape when it lands.04A

Maya

Apply bundle discounts before loyalty in the cart total.

M

01

agent 00124m

12 tool calls, 2 messages

Totals are built in totalFor. bundleDiscount already exists in src/rules.ts and nothing calls it, so the change is one line in the total rather than a new rule.

Maya

Right. Leave the line builder alone, only change the total.

M

01

agent 00118m

Agreed. The cart and its lines both come from Ablo, so the total reads in one screen and I left the rounding untouched.

pricing.ts+4-2

01

agent 00114m

6 tool calls

Ready to write the discount lines now.

Ask, plan, or take the row yourself

Agentagent 001

schema.ts

pricing.ts

checkout.ts

rules.ts

pricing.test.ts

import { defineSchema, model, z } from '@abloatai/ablo/schema';

// One definition per shape. Every agent and every surface

// reads these models, so none of them can drift.

export const schema = defineSchema({

carts: model({

region: z.enum(['us', 'eu', 'apac']),

tier: z.enum(['standard', 'plus']),

status: z.enum(['open', 'priced', 'paid']),

}),

lines: model({

cartId: z.string(),

qty: z.number(),

unit: z.number(),

}),

Give a fleet of coding agents one repository

Each agent holds the file it is changing, queues for the ones it cannot take, and hands it back the moment a person starts typing.

We’re building the collaboration layer for apps where AI agents collaborate with a shared, auditable state.

Ablo is collaboration infrastructure for AI agents

Three ways to reach the same coordinated state

Multiplayer state in TypeScript.

Scoped permission for agents.

Your sync engine, in any AI assistant.