Scitrera Ecosystem Messaging Spec

A universal ChatMessage / ToolCall JSON schema and streaming-event vocabulary for AI agent platforms — a single, transport-agnostic shape for chat messages, tool calls/results, and incremental (streamed) updates, with reference implementations in Go, Python, and TypeScript that are fixture-for-fixture compatible.

It is the wire/at-rest contract used across the Scitrera AI platform, but the schema itself is generic and dependency-light; any agent runtime, frontend, or storage layer can adopt it.

Layout

Path	Purpose
docs/UNIVERSAL_MESSAGE_SPEC.md	Normative spec (v1.0) — the source of truth
go/	Go implementation — module github.com/scitrera/ecosystem-messaging-spec/go
python/	Pydantic v2 implementation — scitrera-messaging-spec (PyPI)
typescript/	TS types + pure reducer — @scitrera/messaging-spec (npm)

All three expose the same schema, the same streaming event vocabulary, and an apply_event / applyEvent reducer with matching fixtures.

Install

Go

go get github.com/scitrera/ecosystem-messaging-spec/go

import spec "github.com/scitrera/ecosystem-messaging-spec/go"

m := spec.NewChatMessage("msg_1", spec.RoleAssistant)
m.Content = []spec.ContentPart{spec.NewTextPart("hello")}

Python

pip install scitrera-messaging-spec

from scitrera_messaging_spec import ChatMessage, TextPart
m = ChatMessage(id="msg_1", role="assistant", content=[TextPart(text="hello")])

TypeScript

npm install @scitrera/messaging-spec

import { makeChatMessage } from "@scitrera/messaging-spec";
const m = makeChatMessage({ id: "msg_1", role: "assistant",
  content: [{ type: "text", text: "hello" }] });

See the per-language READMEs (go/, python/, typescript/) and the normative spec for the full schema and the streaming-event reducer.

Status

v1.0 — stable schema across all three languages. See the spec doc for the normative shape and versioning.

License

Apache-2.0 — see LICENSE and NOTICE.