Back to Delivery Hub
AI-Native

Your Salesforce Org,
Powered by AI

Not a chatbot bolted on — AI woven into every workflow. Delivery Hub ships a full Model Context Protocol server, AI-generated summaries, predictive estimates, and intelligent triage. Bring your own Anthropic API key. No per-seat AI fee.

Model Context Protocol

The MCP Server

Delivery Hub exposes a full MCP server that turns your Salesforce org into a Claude-compatible workspace. Claude can read work items, create tasks, generate documents, run reports — all through the standard MCP protocol. No middleware. No glue code. Just tools.

7 Claude-Compatible Tools

delivery_query_tasks

List and filter work items by status, assignee, workflow, or sprint

delivery_get_task

Get full details for a single work item including logs and comments

delivery_update_stage

Move a work item to a new workflow stage with transition validation

delivery_add_comment

Post a comment on any work item with @mention support

delivery_log_hours

Log time against a work item with date, hours, and description

delivery_dashboard

Pull velocity metrics, capacity data, and sprint summaries

delivery_browse_bounties

List open bounties with claim status and reward amounts

mcp-session

// Claude reads the board

tool: delivery_query_tasks

filter: status=In Progress, sprint=current

// Claude generates a summary

Found 12 items. Generating board summary...

// Claude updates a work item

tool: delivery_update_stage

item: WI-1042 → Code Review

Done. Stage updated, activity logged.

// Claude logs time

tool: delivery_log_hours

item: WI-1042, 2.5h, "Refactored sync engine"

Logged. Total: 14.5h / 16h estimated.

6 AI Features, Zero AI Fee

Every AI feature runs on your own Anthropic API key. Cloud Nimbus never touches your prompts or charges you for AI usage.

AI Board Summary

One-click AI-generated board summary via /api/board-summary. Digests the current sprint state into plain English — blockers, velocity trends, and team load — so your standup takes 2 minutes instead of 20.

AI Work Item Descriptions

Auto-generated work item descriptions from titles and context. Saves 5-10 minutes per item for PMs who hate writing descriptions. One click, full scope and acceptance criteria.

AI Estimation

Predictive effort estimates based on historical work log patterns. Uses logged hours from similar items to suggest sizing — so your estimates are grounded in actual data, not gut feel.

Smart Attention Score

The deliveryScore LWC calculates priority based on age, blocked state, dependency depth, SLA proximity, and assignee load. AI-informed triage that surfaces what actually needs attention.

Work Item Refiner

AI suggests improvements to work item scope, acceptance criteria, and implementation approach. Turns vague one-liners into well-structured, actionable tickets your team can actually build from.

Weekly Digest Generation

Cron-driven AI weekly digest summarizing activity, top clients, actioned vs dismissed counts. Lands in your inbox every Monday so you start the week with full context.

Pricing

Bring Your Own API Key

$0 AI Fee

Delivery Hub never proxies your AI calls through Cloud Nimbus servers. You plug in your own Anthropic API key in Salesforce Custom Metadata (OpenAIConfiguration__mdt), and all AI features run directly against your account. You pay Anthropic directly at their published rates — typically pennies per request. No markup, no per-seat AI tier, no “Enterprise AI” upsell.

Board Summary

~$0.01-0.03 per summary

Auto-Description

~$0.005 per item

Weekly Digest

~$0.02-0.05 per digest

Estimates based on typical Anthropic API pricing. Actual cost depends on prompt length and model selection.

AI-Native Project Management.
Inside Salesforce. Free.

Install Delivery Hub. Add your Anthropic API key. Your board summaries, estimates, and triage scores light up immediately. The MCP server is ready for Claude from day one.