Prospector

Perplexity for design team sponsors — reads your docs, ranks 1,600+ companies by fit, and drafts the outreach.

Waterloo design teams build satellites, solar cars, and autonomous vehicles — projects that rival professional engineering efforts. But finding sponsors to fund them still means team leads manually cold emailing hundreds of companies, on top of a full course load, hoping someone replies.

Prospector fixes that. Point it at your GitHub, Notion, or Confluence and it builds a picture of what your team needs. Ask it for sponsors — it searches 1,600+ companies ranked by technical fit and Waterloo affinity, explains every match, and writes the cold email.

Or view it on YouTube: https://youtu.be/Bt6CjH47xMQ

• Sponsor discovery — semantic search over 1,600+ companies, re-ranked by an LLM for your specific query
• Waterloo-aware — companies that already sponsor design teams rank higher
• Contact finding — scrapes company sites for real emails, falls back to MX-validated suggestions
• Match explanations — specific reasons why each company fits, not generic filler
• Outreach drafting — generates and sends cold emails via Gmail
• Team context ingestion — GitHub, Notion, Confluence, any URL
• RAG chat — ask anything about your team or the sponsor database

pip install -r requirements.txt

1. Go to discord.com/developers/applications → New Application
2. Bot tab → Add Bot → Reset Token → copy it (this is DISCORD_TOKEN)
3. Scroll down → enable Message Content Intent
4. OAuth2 → URL Generator → scopes: bot, applications.commands → permissions: Send Messages, Read Messages, Use Slash Commands
5. Copy the generated URL → open it in browser → add bot to your server
6. Right-click your server name → Copy Server ID (this is GUILD_ID)

cp .env.example .env
# fill in all keys above

Run once, re-run to update.

python scraper/run.py

Sources: Waterloo design team sponsor pages, Velocity incubator, Wikipedia engineering categories, curated seeds.

Output: ~1,600 entities with summaries, tags, and Waterloo affinity evidence stored in Supabase.

bash scripts/run_bot.sh

You'll see [on_ready] ready when it's up. Slash commands sync automatically on startup (~5s).

/setup-team       Register your team — point it at your GitHub org URL
/configure-team   Join the team (every member runs this)
/analyze-team     Load your team context — see blockers, stack, recruiting gaps

/find-sponsors    Search ranked sponsor candidates ("we need RF hardware support")
/explain-match    Drill into why a specific company was recommended
/recruit-gap      See what roles your team is missing

/add-context           Add a Notion page, Confluence space, or URL to your team's knowledge
/remove_from_memory    Delete specific chunks by search query
/nuke                  Wipe all data for your team

/chat           RAG chat — ask anything about your team or sponsors
/sample_email   Draft a cold email to a matched company
/send_email     Send the draft via Gmail

Testing the RAG pipeline without opening Discord:

python test_rag.py "we need RF hardware support"

• Bot: discord.py
• Backend: FastAPI
• Database: Supabase (pgvector)
• Scrapers: httpx, trafilatura, BeautifulSoup4