ys88.io is a private suite of APIs and orchestration primitives for building agent-driven, event-based applications — e-signature, messaging, document intelligence, and workflow automation that AI models can drive end-to-end.
import ys88
client = ys88.Client(api_key="ys88_...")
# Let an AI agent drive an end-to-end e-signature flow
envelope = client.docusign.create_envelope(
template="engagement_letter_v2",
recipients=[{"email": "[email protected]"}],
fields=agent.extract_fields(intake_notes),
)
# Notify the signer by SMS and watch for completion
client.messaging.send(to=phone, body=envelope.signing_url)
client.webhooks.on("envelope.completed", archive_to_vault)A focused set of building blocks covering the most common automation patterns — shaped by real use inside our own operations.
Create, send, and track Docusign envelopes programmatically. Template-based signing, recipient routing, and completion webhooks driven by AI agents.
Send and receive SMS programmatically. Two-way messaging, delivery receipts, and history, with LLM-friendly message threading.
Real-time event delivery with automatic retries, payload signing, and per-endpoint filtering. Fan out to queues or invoke agents directly.
Chain API calls, conditionals, and delays without managing infrastructure. Designed so LLMs can compose and invoke flows at runtime.
Inbound and outbound call handling with programmable call flows, forwarding, and recording — pairs naturally with AI transcription and summarization.
OCR, extraction, and classification for PDFs and scans. Tiered pipeline blends native text layers, pixel OCR, and vision-model anchoring.
Usage dashboards, delivery rates, latency metrics, and audit logs — every agent action recorded and queryable.
Versioned REST APIs plus Model Context Protocol servers so agents can call every capability directly. SDKs for Python, Node.js, and Go.
Scoped API keys, per-workflow credentials, encrypted-at-rest secrets. Built to keep agent-generated calls least-privilege by default.
ys88.io connects directly to the systems our workflows touch. Each integration is wrapped in an LLM-callable interface so AI agents can plan, execute, and verify across services in a single run.
E-signature envelopes created, routed, and tracked via the Docusign eSignature REST API as part of internal automation and AI-driven agreement workflows.
Purpose: Automation & AI tools · Audience: Internal
Account aggregation, transactions, balances, holdings, and statement downloads for consolidated financial observability across institutions.
Claude models orchestrate tool use across every other integration — reading documents, drafting replies, filing paperwork, and triggering signing flows.
Read-only and authoring access for inbox triage, attachment extraction, and filing. Powers agent-driven inbox workflows and document archiving.
Voice and SMS primitives for notifications, two-way conversations, and call-flow automation.
Workers, D1, KV, Queues, and R2 provide the edge runtime, storage, and queue fabric for long-running agent workflows.
Concrete workflows running on ys88.io today — all internal, all driven by AI agents.
An agent drafts an engagement letter from intake notes, routes it via Docusign to the correct signatories, watches for completion, files the executed PDF to Drive, and updates the internal tracker. End-to-end, no human keystrokes between draft and archive.
Incoming email is classified, attachments extracted and OCR'd, statements routed to the right account folder with canonical filenames, and anything requiring action is summarized into a short daily digest.
Nightly sync pulls balances, transactions, holdings, and statements across all linked institutions into a consolidated database. Agents answer "what changed this week?" without touching a spreadsheet.
Large-volume document review, IDR response drafting, and citation anchoring — the same primitives that power agreement automation also handle discovery-scale document sets.
ys88.io is currently deployed as an internal toolkit. Every primitive is shaped by real operational need — signing documents, filing statements, answering inbox questions, and orchestrating the models that do the work. The integration purpose is automation and AI tools, and every API is designed to be invoked by LLM agents as naturally as by a developer.
The private beta is extended selectively to collaborators who want the same primitives for their own agent-driven workflows. External products, public sign-ups, and broad distribution are not in scope today.
ys88.io is in private beta. We're onboarding developers who want to build agent-driven, automation-heavy workflows on top of the same primitives we run internally.
[email protected]