Automation and research for Builders Collective Ottawa sponsor and community outreach.
The repo has three event-specific pieces:
- Local n8n for Ottawa. A self-hosted n8n instance for Ottawa automation. The Grokbot workflows run here. The Snapchat EasyLens workbook does not.
- V1 lead list for the Grokbot event. Two n8n workflows and
BCO_Tech_Sponsor_Lead_List_V1.xlsx. They build a company-level sponsor list from Ottawa public directories and Google Maps. - Toronto sponsor workbook for Snapchat EasyLens. Research notes, scored shortlists, and
Snapchat_EasyLens_Toronto_Sponsor_Leads.xlsxfor the EasyLens × Snapchat workshop on November 12, 2026.
Company-level discovery and sponsor research live here. Sending mail does not. Workflows stop at a sheet or workbook. Email drafts follow BCO_Drafting_Instructions.md, and a send happens only when someone explicitly asks for the Grokbot V1 list.
| Path | What it is |
|---|---|
compose.yaml |
Ottawa n8n service. Binds to 127.0.0.1:5678 and reads secrets from a root .env. |
lead-discovery/ |
Grokbot V1 workflows and sheet template, plus the Snapchat EasyLens research CSVs and workbook builder. V1 setup steps are in lead-discovery/SETUP_README.md. |
research_notes/ |
Source notes behind the Snapchat EasyLens shortlists (AI, company, and creator tracks). |
reports/ |
The three finished Snapchat EasyLens shortlist write-ups. |
Recovered KNBA and Invest Ottawa files/ |
Earlier Kanata North Business Association and Invest Ottawa exports. The workbook builder checks new domains against bco_researched_master.csv and bco_master_leads_checkpoint.csv. |
BCO_Drafting_Instructions.md |
Voice, structure, and send rules for outreach drafts. |
local-files/ |
Host folder mounted into n8n at /files. Gitignored. |
You need Docker and an Apify API token. The Grokbot workflows also need a Google Sheet created from the V1 workbook.
cp lead-discovery/.env.example .envFill in the root .env. Compose also expects these variables, which are not in the example file:
N8N_ENCRYPTION_KEY= # long random string; keep it stable so saved credentials still decrypt
GENERIC_TIMEZONE=America/Toronto
TZ=America/Toronto
APIFY_TOKEN=
BCO_GOOGLE_SHEET_ID=Then:
docker compose up -dOpen http://127.0.0.1:5678. After any .env change, restart the container so n8n picks up the new values. N8N_BLOCK_ENV_ACCESS_IN_NODE=false is set in Compose so workflow expressions can read $env.APIFY_TOKEN and $env.BCO_GOOGLE_SHEET_ID.
Do not commit .env. Tokens and the encryption key stay on the machine that runs n8n.
These workflows run on the Ottawa n8n instance. They only build the Grokbot sponsor list. They do not feed the Snapchat EasyLens workbook.
Full click-by-click setup, including the Google Sheet tabs and credential reconnect, is in lead-discovery/SETUP_README.md.
Import both workflows from lead-discovery/, then reconnect every Google Sheets node to the BCO Google credential. The credential id baked into the JSON is a placeholder.
| Workflow | What a manual run does |
|---|---|
BCO V1 - Directory Sources to Google Sheets |
Scrapes Kanata North BA, Invest Ottawa companies, and the Ottawa Board of Trade technology directory through Apify Web Scraper. Appends raw rows to Raw - Directories, filters for technology relevance, and upserts into Master Leads on Dedup Key. |
BCO V1 - Google Maps Tech Leads to Google Sheets |
Runs six Ottawa searches (software, AI, cybersecurity, cloud, telecommunications, robotics), 20 places each. Appends raw rows to Raw - Google Maps and upserts the same master tab. Reviews, images, place-detail pages, and contact enrichment are off so the first run stays small. |
Upload lead-discovery/BCO_Tech_Sponsor_Lead_List_V1.xlsx to Google Drive and open it as a Sheet first. Leave the tab names as Master Leads, Raw - Directories, and Raw - Google Maps. Put that spreadsheet id in BCO_GOOGLE_SHEET_ID.
After a run, score Master Leads by hand: BCO Relevance, Sponsorship Capacity, Ottawa Presence, and Community Signal. Priority Score is calculated from those. Directory page structure changes; if a source comes back empty, update that source’s page function in the workflow.
These workflows collect public company records. They do not look up employees or personal inboxes.
This track is separate from the Ottawa n8n instance and the Grokbot V1 list. lead-discovery/build_toronto_sponsor_workbook.py turns the lead-discovery/research_*.csv files into lead-discovery/Snapchat_EasyLens_Toronto_Sponsor_Leads.xlsx for the November 12, 2026 EasyLens × Snapchat workshop at the Snapchat Toronto office.
It keeps one row per domain, drops excluded or duplicate domains, scores fit, Toronto presence, capacity, contactability, and speed, and writes four sheets: Sponsor Leads, Methodology, Event Brief, and Excluded & Duplicates. A successful run asserts 75 unique domains, split into tiers A (1–20), B (21–50), and C (51–75).
python3 -m pip install openpyxl
python3 lead-discovery/build_toronto_sponsor_workbook.pyThe research CSVs are the inputs. The Snapchat_EasyLens_Toronto_Sponsor_Leads*.csv files are exports of that workbook for Drive and Google Sheets. Narrative shortlists, each capped at 18 companies with a published company-owned email, are in:
reports/Toronto company sponsors.mdreports/Toronto AI sponsor candidates.mdreports/Toronto creator sponsor candidates.md
Supporting notes sit in the matching folders under research_notes/.
BCO_Drafting_Instructions.md is the source of truth for voice, links, personalization, and which list may be sent.
Keep the lists separate:
- BCO Master Leads 2.0 is draft-only. Create Gmail drafts. Do not send.
- Grokbot V1 list (
BCO_Tech_Sponsor_Lead_List_V1) is drafted the same way, and sent only when explicitly asked. - Snapchat EasyLens uses the Toronto workbook. Draft from that list only when asked. It is not the Grokbot sheet.
Before drafting a batch, check Gmail sent history for each recipient. Row order and the Drafted? column are not reliable; that column is not written back.