TIMBAR / proposal

Prepared for TIMBAR · 9 October 2026

The right fit.
Before they buy.

A reliable saw compatibility database, a helpful website bot and a follow-up system that turns missing information into answers.

See the work & pricing

01 / The outcome

One clear answer.
Evidence behind it.

Build the knowledge first. Put a useful conversation in front of it. Keep improving the answers as TIMBAR learns more.

01

Your compatibility database

Import the initial 88-model spreadsheet, then organise brands, model aliases, years, chain pitch and TIMBAR-approved fit criteria. Attach manuals, replacement parts, photos and available CAD links to the records.

Every record keeps its source, trust rating, agreement or conflict between sources, review date and unanswered questions. Missing information stays visible.

02

A shopper compatibility bot

Ask a few plain-language questions. Help people identify their saw using approved database photos and year ranges, even when they cannot find a model label. Offer photo submission for staff review.

Show Proven compatible, Likely compatible or Not compatible only when the approved rules support the answer. Otherwise say Needs review and hand the case to TIMBAR.

03

A research & follow-up bot

Look for missing specifications and draft specific questions for published manufacturer business contacts. Capture replies, queue new evidence for review and show who the team is waiting on, for what and for how long.

Start with public-web research. Activate individual email follow-ups only after the sender and wording are approved, with a stop rule and a human review route.

04

A team that can run it

A staff dashboard to review answers and unresolved cases, demand reporting for models shoppers ask about, and an opt-in quarter-inch waitlist.

Onboard up to five staff, configure an internal assistant to find approved TIMBAR records and open questions, and hand over the data, source code and operating guide. Train one TIMBAR administrator and the initial team.

The answer comes from evidence

TIMBAR supplies and signs off the fit criteria and what counts as “proven” or “likely.” The AI helps with questions and research; it does not certify fit from a photograph. Conflicting, incomplete or unsupported records go to a person. “Likely” displays the proposed 30-day return promise only after TIMBAR approves the policy and exact wording.

02 / The investment

Four phases.
One hourly rate.

Every line includes its hours and cost at C$125/hour. Low assumes clean inputs and prompt decisions; expected allows normal research and review; high allows data cleanup, source conflicts and integration rework within the agreed scope.

01 / Build the compatibility database

TIMBAR can review the initial model list, the evidence behind each record, and the questions still open.

Hours and CAD cost by estimate scenario
Work item Low Expected High
Confirm fit criteria, priorities and acceptance examples 6 hC$750 8 hC$1,000 12 hC$1,500
Database design and import of the initial 88-model spreadsheet 12 hC$1,500 18 hC$2,250 24 hC$3,000
Initial public-web research: manuals, parts, photos, available CAD and source trust 16 hC$2,000 24 hC$3,000 36 hC$4,500
Staff review screens, evidence history, conflicts and missing-spec queue 8 hC$1,000 12 hC$1,500 18 hC$2,250
Phase subtotal 42 hC$5,250 62 hC$7,750 90 hC$11,250

02 / Give shoppers a helpful fit checker

The website checker answers from approved evidence, collects demand and routes unresolved cases to TIMBAR.

Hours and CAD cost by estimate scenario
Work item Low Expected High
Compatibility rules, evidence grades and human escalation 10 hC$1,250 16 hC$2,000 22 hC$2,750
Chat follow-ups, model/photo selection and optional photo submission 16 hC$2,000 24 hC$3,000 34 hC$4,250
Website integration, quarter-inch waitlist, consent and demand reporting 8 hC$1,000 12 hC$1,500 18 hC$2,250
Phase subtotal 34 hC$4,250 52 hC$6,500 74 hC$9,250

03 / Close the gaps with manufacturers

Approved questions and replies flow into the database; the team can see who owes an answer and how long it has been waiting.

Hours and CAD cost by estimate scenario
Work item Low Expected High
One approved sender/inbox connection and message approval workflow 6 hC$750 10 hC$1,250 14 hC$1,750
Specific question drafts, bounded follow-ups, reply capture and review 12 hC$1,500 18 hC$2,250 28 hC$3,500
Waiting-on dashboard, scheduled refresh and initial qualitative market snapshot 8 hC$1,000 12 hC$1,500 18 hC$2,250
Phase subtotal 26 hC$3,250 40 hC$5,000 60 hC$7,500

04 / Test, launch and hand it over

TIMBAR accepts the release, can manage its data and has a documented route for support.

Hours and CAD cost by estimate scenario
Work item Low Expected High
End-to-end tests: fit outcomes, unknowns, photo flow, access and mobile use 10 hC$1,250 16 hC$2,000 24 hC$3,000
Deployment, usage limits, monitoring, backup and restore rehearsal 8 hC$1,000 12 hC$1,500 18 hC$2,250
Onboard up to five staff and configure a scoped internal data assistant 6 hC$750 10 hC$1,250 14 hC$1,750
Source/data handover, written operating guide and one training session 4 hC$500 6 hC$750 8 hC$1,000
14-day post-handover support window: triage and agreed-scope fixes 6 hC$750 8 hC$1,000 12 hC$1,500
Phase subtotal 34 hC$4,250 52 hC$6,500 76 hC$9,500
Total implementation, testing, handover & support
Project total Low Expected High
All four phases 136 hC$17,000 206 hC$25,750 300 hC$37,500

These are effort estimates, not a fixed-price contract. Work is billed for actual recorded hours. Weekly updates show time used, decisions due and forecast to complete. Additional scope or a forecast above the high estimate requires written approval before proceeding.

Included

  • Initial import and research pass, review screens and one website integration.
  • Shopper checker, approved email workflow, demand reporting and staff assistant setup.
  • Mobile and desktop testing, access checks, deployment, spend limits and a backup/restore rehearsal.
  • Code and data handover, one training session and 14 calendar days of post-handover support. Support effort is included in the estimates; business-day triage and agreed-scope fixes, without a 24/7 response guarantee.

Excluded / separately scoped

  • AR, scanning, automatic dimensions or fit certification from photos; hardware redesign or physical lab testing.
  • A new general-purpose team platform, live meeting notetaker, native apps, remote control, billing or a product rename.
  • Paid market-share datasets, reliable unit-sales estimates, purchased CAD licences and unrestricted ongoing research.
  • Unlimited manufacturer outreach, additional inboxes or website integrations, ecommerce checkout changes, new languages and a private-server deployment.
  • Third-party bills, taxes and ongoing maintenance after the support window.

03 / The schedule

Useful early.
Complete in 6–8 weeks.

Expected delivery is six weeks, with up to two additional weeks for data cleanup, integration issues and acceptance. This assumes a booked delivery slot of roughly 35–40 project hours per week and TIMBAR decisions within two business days. The 14-day support window follows handover.

  1. Before kickoff

    Make the inputs ready

    TIMBAR provides: Alex’s spreadsheet; Tisha’s criteria document; photos with and without housing, including the agreed reference-coin views; priority models; one decision-maker and support contact; website/staging access; an approved inbox owner; and a company-funded AI account.

    We agree: evidence labels, initial research limits, customer handoff, photo consent, retention and the return-policy wording. Confirm the start date and approve the deposit.

  2. Week 1

    Import, organise and expose the gaps

    Seed the database, demonstrate a small reviewed model set, start public-web research and agree a prioritised unknown-spec list.

    TIMBAR provides: corrections to aliases and priorities, reviewed examples of all compatibility outcomes, and confirmation of photo rights.

  3. Week 2

    Review the evidence together

    Deliver the staff database screens, source trust/conflict fields, initial research pass and missing-spec queue. First acceptance milestone: database ready to review, including unresolved records.

    TIMBAR provides: approval of evidence rules and the supported model subset, website placement and a named human escalation owner.

  4. Week 3

    Put the shopper checker in staging

    Demo follow-up questions, photo selection/submission, all outcome labels, human handoff, demand reporting and the quarter-inch waitlist. Test on the actual website integration.

    TIMBAR provides: representative shopper questions, approved customer wording and test feedback.

  5. Week 4

    Connect approved manufacturer follow-ups

    Connect the sender and inbox, enable approved individual questions and bounded replies, and show waiting time and response history. Schedule a refresh and deliver the qualitative market snapshot.

    TIMBAR provides: sender/domain approval, Tisha’s message approval, initial contact list approval and a manual follow-up owner.

  6. Week 5

    Rehearse the full experience

    Test fit decisions, unknowns, permissions, consent, mobile use, inbox handling, spend controls and restore. Set up the scoped internal assistant and rehearse first use with the initial staff.

    TIMBAR provides: staff availability, acceptance feedback and a decision on which reviewed records may go live.

  7. Week 6

    Accept, release and hand over

    Complete acceptance, publish the approved checker, train the team and transfer code, data, configuration and the operating guide. Begin the 14-day support window.

    TIMBAR provides: release approval, final website access and the person who will own daily review and customer escalation.

  8. Weeks 7–8
    if needed

    Resolve the remaining delivery risks

    Allow for source conflicts, data corrections, integration changes or acceptance fixes within the high estimate. Manufacturer replies may arrive later; the system ships with honest unknowns rather than waiting for every answer.

October 20 and the October 29 product launch

October 20 is treated as a possible limited internal preview, subject to immediate kickoff, clean inputs and availability. The notes do not settle which deliverable is due that day. It is not a promise of the full system. For October 29, a checker limited to reviewed models with a human handoff may be considered only after TIMBAR accepts the evidence and website tests pass. Confirm both targets at kickoff; the full delivery estimate remains 6–8 weeks.

04 / Scope assumptions

Clear boundaries.
Fewer surprises.

The first release

  • One TIMBAR website, English, one sender/inbox and up to five staff. A standard embed and inbox connection; custom ecommerce or legacy integrations need a separate estimate.
  • Import up to 88 initial model records. Budget one research pass across three priority retailer/platform sources and two parts/manual sources; record additional discoveries for prioritisation. Importing a record does not mean fit is verified.
  • Initially up to 10 manufacturers and 50 individual published business contacts, with an agreed follow-up cadence and stop limit. No manufacturer response-time guarantee.
  • Initial refresh scheduling and a qualitative brand/model/price landscape are included. Verified sales or market share require separately approved data and work.

Decisions and responsibilities

  • The internal assistant configuration assumes an existing working team workspace or a simple data-search interface. A broader personal-assistant service, platform rebuild and live notetaker are outside this estimate.
  • TIMBAR decides technical fit, approves claims, owns the return promise and handles customer escalations. The quarter-inch waitlist signals demand; it does not promise a product or delivery date.
  • Only approved data and customer notices go live. Agree image permissions, photo handling, data retention and AI-provider settings before customer use; no blanket training or hosting-location promise is made here.
  • Delayed inputs or approvals move calendar milestones. Extra models, sources, workflows or support are change requests at C$125/hour, agreed in writing.

05 / After launch

A small operating budget.
Room to grow.

A planning allowance for about 1,000 shopper conversations per month, an 88-model dataset, up to 20 GB of photos/manuals and modest research refreshes. TIMBAR pays providers directly. These are estimated operating costs, separate from implementation.

Estimated third-party costs / month · CAD
Service Monthly allowance
Hosting, database and photo storage C$7–C$25
AI usage allowance C$25–C$125
Sender mailbox allowance C$0–C$20
Total, excluding maintenance and tax C$32–C$170

Hosting & storage · C$7–C$25/month

For estimating, a small Cloudflare-hosted setup uses static Pages, a Workers API, D1 records and R2 photos. Static asset requests are free. The Workers paid base converts to about C$7/month; its included usage and D1’s included 5 GB should cover this modest starting workload. R2 Standard includes 10 GB-month; the next 10 GB is about C$0.21/month for storage, before operation charges. The upper allowance leaves room for usage and monitoring.

Official pricing: Pages · Workers · D1 · R2

AI usage · C$25–C$125/month

TIMBAR supplies its own OpenRouter account. A low-cost reference model, Gemini 2.5 Flash Lite, currently converts to C$0.14 per million input tokens and C$0.56 per million output tokens. At 20,000 input and 2,000 output tokens per complete shopper conversation, 1,000 conversations use about C$3.92 of inference, or roughly C$4.14 including the Standard 5.5% platform fee allocation, before any minimum credit-purchase fee or tax.

The larger allowance covers research, document/photo processing, retries and occasional stronger models. Model choice is confirmed against actual answer tests; the low-cost example is not a quality guarantee. Agree a usage ceiling and pause AI work or hand off to staff at the limit. Higher traffic and recurring research increase cost.

Official pricing: Reference model · OpenRouter fees · Billing details

Email · C$0–C$20/month allowance

Use an existing TIMBAR business mailbox where possible, with no additional mailbox charge assumed. C$20 is a provisional reserve if a dedicated sender or delivery service is needed, not a quoted vendor price. Select the provider and approve its bill before activation. Bulk marketing and paid research subscriptions are excluded.

Optional maintenance · C$500 or C$1,000/month

4 hours / C$500 for routine checks, data corrections and small fixes; or 8 hours / C$1,000 for more frequent updates and improvements. Either option is month-to-month, with priorities agreed each month. Unused hours do not roll over. Extra work is C$125/hour with approval. No retainer is required to accept the project.

Pricing checked 9 October 2026. Every amount on this proposal is CAD. Vendor pages may quote USD; conversions use a budgeting assumption of C$1.40 per US dollar, not a current exchange-rate quote. Actual invoices depend on exchange rates, provider plan, account-wide included usage, tax and measured consumption. Final hosting and provider choices are confirmed at kickoff.

06 / Working together

Milestones you can review.
A rate you can plan around.

Suggested payment schedule against the expected C$25,750 estimate:

  1. Kickoff deposit30% · C$7,725
  2. Database accepted30% · C$7,725
  3. Checker and follow-up system accepted30% · C$7,725
  4. Final handover10% · C$2,575

Suggested commercial terms

30% deposit to reserve the delivery slot. Milestone invoices reflect actual recorded hours and deposit credit; the amounts above illustrate the expected budget and are reconciled at final handover. Suggested invoice terms: 14 days. Applicable taxes and approved provider charges are additional.

The estimate remains open for 14 days from this proposal’s date. Written acceptance confirms the scope, start date, payment schedule, supported records and the named acceptance owner.

Changes & handover

Changes are quoted in hours at C$125/hour before work begins. We explain their cost and effect on the schedule, then wait for written approval. The high estimate is the approval ceiling for this agreed scope, not permission to add features.

On payment, TIMBAR receives the project-specific source code, data exports, deployment configuration and operating guide. Third-party licences and any pre-existing tools remain subject to their own terms. Provider accounts should be controlled by TIMBAR.

Next step

Approve the scope. Send the inputs. Pick the start date.

Eric will confirm the first supported models, the October preview targets and the delivery slot with TIMBAR before kickoff.