Skip to content
View packages

GoSeedly Handyman Kit / INSTALLATION

A clear path
to your workspace.

Bring the kit into your compatible Seedly installation, preserve existing work, and choose the starter content that fits your team.

01

Start with your Seedly project.

Use your separately obtained Seedly CRM 5.8.4 installation with extension API 1, a configured backend, and working authentication. Keep a private source checkpoint, including existing customizations. Use Node.js 20 or newer.

02

Run guided setup.

Extract the kit outside the Seedly source folder. From the kit directory, run node setup.mjs. The check identifies supported integration files, changed files, and destination conflicts before installation.

03

Build and connect.

Deploy the included required helper from your authoritative backend source before connected edits. Follow QUICKSTART.md with your project’s existing package manager. Build the web app, sign in, and open your existing /location/YOUR_LOCATION_ID address. Verify access, navigation, and your important CRM flows.

Install the interface.
Choose the starting content.

Frontend installation and optional account content are separate. The included original job helper is required for connected creation, editing and starter setup. It preserves native account permissions and unrelated custom fields, re-checks the quarter hour rule along with the handyman, request date and recorded-visit rules, and compares the current record state before saving. It exposes no stage mutation at all, so pipeline stages stay in your CRM’s own views under your existing roles. Deploy the helper from your authoritative backend source before connected edits. Pipeline presets and workflow drafts remain optional. Follow QUICKSTART.md and STARTER_CONTENT.md for the exact integration. The frontend installer does not deploy your backend.

Pick your starting structure.

Starter setup offers Residential small jobs, Property manager work orders, Recurring maintenance visits, Larger project quotes. Select individual pipelines and drafts, or use a suitable existing pipeline with the exact stage needed by the chosen draft.

Adapt each workflow.

All nine workflows remain inactive drafts. Review their native triggers, internal tasks, notes, timing, and assignments before publishing in Seedly. Timed waits do not check whether a task is complete or the stage has changed. These examples do not send customer messages.

Explore sample mode.

Explore nine fictional jobs planned across the five working days between two handyman labels, with thirty-one task rows and their hour estimates, eleven material rows of which four are not recorded on hand, and five recorded visits. One of the nine runs to 7.75 estimated hours, so the rose bar and the over six hours flag are visible without editing anything, and eight task rows are recorded done so the ticks and the struck labels both appear. Every sample date sits inside the window the stack always draws, whatever weekday you open it on. Example edits stay in the current page’s memory and reset on navigation or reload; sample jobs are never inserted into your CRM. Add ?example=1 to the home or custom-view URL. Leave sample mode to return to your real account.

INSTALLING WITH AN AI AGENT?

Give it the context.
Keep your custom work.

Existing records and unrelated custom source do not require starting over. Modified integration files need a deliberate merge.

Start with the copy-and-paste prompt in INSTALL_WITH_AI.md.
Read AGENT_INSTALL.md for routes, permissions, provider wiring, backend contracts, and compatibility requirements.
Run node setup.mjs check "/path/to/seedly" before modifying files.
Preserve existing extensions, providers, account controls, custom pages, and unrelated data.
Use a reviewed merge for changed integration files; never replace them with clean copies or bypass the installer’s fingerprints.
Deploy the required write helper from the authoritative backend source before connected edits; preserve existing functions and authorization.
Build and verify account access, original CRM modules, new screens, saving, and reload behavior before publishing.

Make the workspace
feel like your team.

Edit apps/web/components/devlaunch-handyman/config.ts to set company name, descriptor, promise line, and default pipeline name, then rebuild. The design applies to the entire web installation. Use separate web installations to evaluate different full-interface kits, rather than stacking their installers.

Starter removal, updates, and rollback

Review unused starter additions for removal in Starter setup before removing the frontend. Server checks preserve edited or published workflows and pipelines with jobs or references. The process targets recorded kit additions rather than sweeping account records.

Preserve your local configuration and source edits. Preview frontend removal with node setup.mjs remove "/path/to/seedly" --dry-run. The automatic uninstaller uses a private restoration receipt and refuses to overwrite later edits. Keep that receipt private and out of public hosting or source exports. Removing frontend files does not remove account data or undeploy backend functions.

For an update, follow the documented removal/install process and reapply your preserved changes. A manually merged installation needs its own private change ledger and reviewed source rollback.

What the workspace records

A job is a native contact-linked opportunity with a namespaced service address label, request summary, handyman label, recorded request and planned visit dates, priority, job notes, an immutable supported currency, and three archive-only child collections with stable IDs: task rows with a quarter hour estimate and a done flag, material rows with a whole quantity and an on hand flag, and visit rows with a date, a handyman label, recorded actual hours, an outcome and a note. The recorded quoted reference stays separate from native opportunity value. Other custom fields are preserved.

Every hour figure, both a task estimate and a recorded actual, must be a quarter hour from 0.25 to 24. A value that is not a multiple of a quarter hour is refused rather than rounded, because rounding somebody’s estimate is editing it, and quarter hours are exactly representable so a card total never drifts. Nothing multiplies an hour by anything: there is no hourly rate, labor rate table, price book, markup, margin or pricing advice anywhere in the kit. A job that has not been itemized reports No tasks recorded rather than zero hours and draws no bar at all. Evidence-based rules are identical on create and on save and re-checked by the server: a recorded handyman requires a recorded visit date, a visit date may not precede the recorded request date, and a task can only be recorded done once the job records a visit, because a punch list cannot tick itself. Three non-rules are deliberate: a visit log date is not constrained against the planned date because work slips and work gets pulled forward, two visit rows may share a date, and recorded actual hours are never constrained against the estimate. A recorded visit date and handyman label reserve nobody and create no calendar appointment; there is no scheduling engine, availability check, capacity model or calendar sync. A material’s on hand flag is a note about the van, not a stock level, and the kit orders nothing from anybody. There is no route optimization, GPS or drive time estimate, no invoicing or payments, no customer messaging of any kind, and no timekeeping or payroll. Totals are grouped by currency and never added across currencies. The pipeline stage stays your CRM’s own: the helper exposes no stage mutation at all, so ticking a task never moves a card and moving a card never ticks a task.

The kit’s original operational screens share the frame with full CRM modules from your own Seedly source. Existing account permissions, feature gates, and configured providers still determine access to those modules.