it sees what you can't.

parasol.

the trading agent that protects you from solana — finds the 2% of tokens that aren't scams, trades them for you, and exits before they rug.

live on mainnet paying users colosseum frontier 2026

parasol.so · @ParasolSolana

01 — one-liner

team · founder-market-fit

built by a trader who got rugged one too many times

Daniel Stephenson

solo founder. full-stack. trades solana memecoins daily for two years.

  • previously built MoltyDEX — solana dex aggregator on jupiter, live at moltydex.com
  • previously built Deepwave — perp trading calculators, ranks #1 organic for "funding rate calculator"
  • built parasol's entire stack: trading engine, ai pipeline, brand, content, growth, infra
  • shipping daily — 40+ commits in the last 48 hours

unfair advantages

  • three live solana products — i've been on this exact stack for 18 months
  • months of live trading data — every trade trains the ml learner. competitors start from zero
  • trader by trade, builder by craft — i ship the thing i need, then sell it to people like me
  • no investor pressure — 100% bootstrapped, can move at hackathon pace

"the founder who lives the problem ships the right product." — me, learning the hard way.

02 — team

traction · real signals

live product. real users. real fees.

live
on solana mainnet with real SOL trades
100+
trades executed by the agent in production
SOL
in platform fees collected directly to PARASOL_FEE_WALLET

what real traders said

"the manipulation shield rejecting tokens with bad holder distribution before i see them is the killer feature. i lose 60% of my trades to rugs i should've spotted."
— sample quote · replace with verbatim from interview log
"i pay $100/mo for axiom + $50 for telegram bots. i'd switch to one tool that does scanning and exits properly."
— sample quote · replace with verbatim from interview log
"i never sell at the right time. if your ai really manages exits, that alone is worth it."
— sample quote · replace with verbatim from interview log

target by submission: 20 trader interviews · 5 verbatim quotes · 10+ active beta testers · 47 blog posts driving organic SEO

03 — traction

hackathon sprint · what we shipped

4 weeks. production-grade additions.

Parasol existed before the hackathon as an autonomous trading agent. During the Frontier window we built the safety intelligence layer that makes Parasol useful to any AI agent on Solana — not just our own.

phantom mcp
parasol_safe_trade_check tool
one MCP call returns rug score, honeypot verdict, holder analysis, slippage estimate. compatible with phantom's wallet MCP.
open source
parasol-sdk under MIT license
standalone TypeScript MCP client + full docs at github.com/Djtrixuk/parasol-sdk. other devs can build on parasol's intelligence layer.
execution
hardened buy/sell cascade
cross-instance buy mutex via redis, dynamic jito tips, blockhash-expired retry, simulate-on-optimised-tx, account-keyed priority fee cache. ~2x faster trades, 5x more reliable.
billing
PCU credit system + api keys
3-tier api keys (free / pro / enterprise) with scope enforcement. PCU deduction on every MCP call. monetises the SDK from day one.
tools moat
3 free trading calculators
position sizing, funding rate, liquidation price. live coingecko + hyperliquid feeds. organic SEO funnel.
learning
whatLearned
three weeks in: trading agents need a safety layer as much as a strategy. axiom is execution. gmgn is alerts. parasol is the only product combining safety intelligence + autonomous execution + open SDK.

see commits: github.com/ParasolSolana/ParasolSolana · sdk: github.com/Djtrixuk/parasol-sdk

04 — hackathon sprint

demo · working product

watch parasol find, filter, trade, exit

[embed demo video here — 2:45 walk-through of three real flows]

full demo: parasol.so/demo · live product: parasol.so

flow 1 · scan + trade

agent finds a new pump.fun token, runs the shield, scores via momentum strategy, executes 0.05 SOL buy in < 2s.

flow 2 · shield rejects

token with 73% top-10 holders + unrevoked mint authority returns AVOID. user never sees it. that's the moat.

flow 3 · ai exit

position +42% — model says take 50% off, trail the rest at 3%. one click. half locked, half running.

05 — demo

business model · how we make money

three revenue streams. live from day one.

per-trade fee

~1% platform fee on every agent trade. routes through our PARASOL_FEE_WALLET. live and collecting today. scales linearly with active users.

PCU credits (SDK)

third-party AI agents pay per call to parasol_safe_trade_check. tiered api keys (free / pro / enterprise). recurring revenue from devs building agent products.

premium tier (planned)

subscription for power users — higher position limits, priority execution, advanced agents (claudia pro, grok pro). free tier remains for paper trading + tools.

unit economics

  • average trader does ~30 trades/week in active periods
  • at 1% fee, ~$0.50 per trade on average position size
  • = ~$60/month per active trader, organically
  • 5-tier referral programme: 50% of platform fees recycled to referrers — zero CAC at scale

why this works: traders already pay $100-300/mo across axiom, photon, gmgn, telegram bots. parasol replaces the stack with one product that also trades for them. retention is locked-in by the data flywheel — switching costs grow with every trade.

06 — business model

opportunity · the market

solana memecoins are a $40B liquid market, growing

$2.5B+
solana DEX volume per day
7M+
tokens created on pump.fun lifetime
2M+
solana active wallets daily

the wedge — memecoin traders

retail traders are overwhelmed. thousands of tokens daily, most are scams. they currently spend $100-300/mo on a fragmented stack of scanners, alerts, and bots — none of which actually trade for them.

the expansion — every AI agent on solana

phantom, openrouter, and 200+ projects are building AI agents that need wallets. none have safety intelligence built in. parasol's SDK is the on-chain judgment layer they need to ship to production.

07 — opportunity

competitors · who else is in this space

parasol is the only product combining all four

Parasol Axiom Photon GMGN
autonomous AI execution full pipeline ✗ manual ✗ manual ✗ alerts only
AI exit management stops + trailing + TP + AI review basic stops basic stops ✗
6-layer manipulation shield 82% reject rate basic basic basic
multi-strategy ML scoring 7 strategies, adaptive weights ✗ ✗ 1 (copy trade)
open SDK / MCP integration MCP + MIT SDK ✗ ✗ ✗
non-custodial wallets ✓ Turnkey MPC ✗ connected ✗ connected ✗ connected
24/7 autonomous ✓ ✗ ✗ ✗

the moat isn't any single row. it's all seven, in one product, that other AI agents can call via MCP.

08 — competitors

gtm · why now

three megatrends meeting in the same quarter

1 · ai agents need wallets

phantom shipped MCP wallet support in march. cursor + claude both ship MCP. agents can finally execute on-chain. but they need judgment — that's parasol.

2 · solana memecoin maturation

$2.5B daily DEX volume + pump.fun's 7M tokens means scams are industrialised. retail traders are at peak pain. a tool that actually filters wins now.

3 · open SDK economics

every AI agent product needs the same primitives. the team that becomes the safety standard captures the market — like stripe did for payments.

go-to-market

  • phase 1 (now) — direct to memecoin traders via X, discord, referral programme
  • phase 2 (Q3) — SDK distribution to MCP ecosystem (phantom, cursor, claude desktop)
  • phase 3 (Q4) — institutional onboarding via white-label agent infra
  • organic SEO — 47 blog posts + 3 free calculators rank for "position sizing", "funding rate", "liquidation"
  • referral programme — 5-tier, 50% revenue share, 3 levels deep. zero CAC engine
  • developer pull — open SDK + tutorials = inbound from agent builders
09 — gtm · why now

roadmap

where we're going

Q1 2026 — shipped

core engine. 3 AI agents (sienna, claudia, grok). live trading via turnkey MPC. multi-wallet. 47 SEO blog posts. 5-tier referrals. fee collection live.

Q2 2026 — just shipped (hackathon)

MCP server with parasol_safe_trade_check. open-source SDK under MIT. PCU billing system. hardened execution cascade. phantom integration guide.

next 6 months — growth

arena (bot-vs-bot competition). premium subscription tier. mobile app. social trading (copy top agents). expand SDK to support cursor/claude desktop natively.

18 months — expansion

cross-chain (base, ethereum L2s). API/SDK partnerships with phantom, jupiter, raydium. multi-agent orchestration. white-label agent infrastructure for institutions.

5 years — vision

parasol becomes the bloomberg terminal for autonomous agents. every AI on-chain calls our intelligence layer before it transacts. category-defining safety standard.

10 — roadmap

risks · what could kill this

the four risks we worry about (and how we handle them)

regulatory uncertainty around ai trading agents
mitigation: non-custodial by design — we never touch user funds. turnkey MPC means users sign their own transactions. paper trading default. clear disclaimers. exposure is < brokerage equivalent.
solana memecoin market crash / volume collapse
mitigation: SDK + MCP layer means parasol can serve any AI agent on solana, not just memecoin traders. cross-chain expansion in roadmap. tools/calculators are SEO assets that survive any cycle.
large incumbent (axiom, jupiter) ships AI features
mitigation: 18 months of live trading data is the moat — they'd need a year to catch up on the ML weights alone. our open SDK means any agent product they build has parasol as a natural component, not a competitor.
solo founder bandwidth
mitigation: shipping pace proves throughput (40+ commits in 48 hours during hackathon). first hire post-funding is a senior engineer. founder time is highly leveraged via AI-assisted dev workflow.
11 — risks

next steps · what we want from you

three asks. in priority order.

1 · users

try parasol. parasol.so. tell us what's broken, what's missing, what makes you stay. we onboard the first 100 testers personally — that's how we keep founder-market-fit tight.

2 · ecosystem partners

if you build wallets, agent frameworks, or AI tooling on solana — let's integrate. parasol's SDK is MIT and the API is documented. github.com/Djtrixuk/parasol-sdk.

3 · seed funding

raising $1.5M to hire 2 engineers, scale the referral programme, ship mobile, and expand the SDK. revenue is live from day one — capital accelerates, doesn't enable.

most-wanted next 30 days:

100 active testers

via direct outreach + colosseum exposure

3 SDK integrations

phantom + 2 agent-framework partners

first $1k MRR

from per-trade fees + early SDK calls

12 — next steps
it sees what you can't.

parasol.

the market is chaos.
parasol is the shade.

live product

parasol.so

demo video

parasol.so/demo

open-source SDK

github.com/Djtrixuk/parasol-sdk

x / twitter

@ParasolSolana

colosseum frontier hackathon · 2026

13 — contact