On an average day, somewhere between 5,000 and 15,000 new tokens launch on Solana through PumpFun alone. Add in tokens from other launchpads and direct deployments, and the number gets even crazier. Nobody can evaluate all of them. Nobody should try.
The question isn’t “how do I see more tokens?” It’s “how do I see fewer tokens, but better ones?”
Over about eight months of trading, I’ve built a screening workflow that takes that ocean of launches and narrows it down to roughly 3-5 trades per day. Some days zero. Rarely more than five. The workflow isn’t perfect — I still take losing trades — but it’s consistent and repeatable, which matters more than perfection in this market.
Here’s the entire process, step by step.
Layer 1: The Automatic Filter
This is the first pass, and it happens before I even look at anything. My token screener is configured with baseline filters that automatically hide tokens that don’t meet minimum criteria.
My current baseline filters:
- Enriched tokens only. This means the token has been through a safety analysis pipeline. Raw, unanalyzed tokens don’t show up at all.
- Graduated to Raydium. I almost never trade tokens still on the PumpFun bonding curve. Graduation means someone cared enough to push it through, which is a basic quality signal.
- Minimum 30 holders. Below this, the token is too early and too risky for my style. Some traders prefer to get in earlier — that’s fine, it’s just not my approach.
These three filters alone eliminate probably 97% of all token launches. From 10,000+ daily launches, I’m now looking at maybe 200-300 tokens that made it through this first gate.
I know that seems aggressive. You might worry about missing winners. And yes, I do miss some. But the winners I miss are overwhelmingly the ultra-early, ultra-risky plays where timing is measured in seconds and success requires luck more than skill. The ones I catch tend to be slightly later entries with much higher probability.
Layer 2: The Visual Scan
Once I open my screener, I’m looking at a list of 200-300 tokens that passed the automatic filters. I scan this list quickly — maybe 2-3 minutes — and I’m looking for specific things.
What catches my eye
- Holder count that’s noticeably higher than similar-aged tokens. If most tokens from the last hour have 40-80 holders but one has 200+, that’s unusual and worth investigating.
- Safety badge showing “Safe.” I generally skip “Warning” and always skip “Danger” tokens. Life’s too short and there are enough safe tokens to trade.
- A name or image that connects to a recognizable narrative. This is subjective, but memecoins are driven by culture. A token that taps into a trending meme, a current event, or an established crypto narrative has a structural advantage over a random name generator output.
What makes me scroll past
- Suspiciously high market cap with few holders. A $500K market cap with 35 holders means a small number of wallets are inflating the price. That’s a setup for a dump.
- Generic or copycat names. If there are five tokens named variations of the same trending meme, most of them are cash grabs riding someone else’s narrative. I look for the one that’s original or first.
- No image or broken metadata. Serious token creators upload an image, write a description, and set up social links. If a token has none of these, the creator put in zero effort, and that usually shows in the outcome.
After this visual scan, I’m down to maybe 10-20 tokens that I want to look at more closely.
Layer 3: The Detail Check
This is where I actually click into individual tokens and spend 30-60 seconds on each one. I have a specific order I check things in, and I’ll bail at any step if something looks wrong.
Step 1: Holder distribution (10 seconds)
What does the top holder breakdown look like? If the top wallet holds more than 8-10% (excluding known addresses like the Raydium pool), I’m usually out. If the top 10 wallets collectively hold more than 40%, I’m definitely out.
What I want to see: a relatively flat distribution where no single wallet dominates. This suggests broad, organic ownership rather than a few insiders controlling the supply.
Step 2: Safety details (10 seconds)
Mint authority revoked? Freeze authority revoked? These are non-negotiable for me. If either authority is still active, the creator can mint new tokens or freeze your wallet. Instant pass.
Step 3: Price chart (15 seconds)
I’m not doing technical analysis here. I’m asking one simple question: is the price currently in a healthy range relative to its history? Specifically:
- If the token just did a 10x in the last 30 minutes, I’m late. Pass.
- If the token is pulling back from a high but still has higher lows forming, that’s interesting.
- If the token is slowly bleeding with declining volume, it’s dying. Pass.
Step 4: Social presence (15 seconds)
Quick Twitter search for the token name or contract. Is anyone talking about it organically? Are there real accounts discussing it, or just bot armies? If there’s zero social presence, the token has no marketing engine and will probably fade.
I also check if there’s a Telegram or Twitter linked from the token’s metadata. Active communities with real conversation are a strong positive signal.
Step 5: Gut check (5 seconds)
After all the data checks, I ask myself one question: would I be comfortable holding this for 4-6 hours if it goes flat? If the answer is no — if I’d panic and sell the moment it dips 20% — then my conviction isn’t high enough and I pass.
After Layer 3, I’m down to maybe 3-5 tokens per day that I feel good about trading. Some days it’s zero, and that’s fine. No-trade days are better than forced-trade days.
Layer 4: Position Sizing and Entry
The tokens that survive all three layers get actual money. But not all equally.
I use a simple three-tier sizing system:
| Conviction Level | Position Size | What Qualifies |
|---|---|---|
| Standard | 0.5-1 SOL | Passes all checks but nothing exceptional. Testing the waters. |
| High | 1.5-3 SOL | Strong narrative, excellent holder distribution, growing social buzz. |
| Max | 3-5 SOL | Everything aligns perfectly. Maybe once a week. |
Most of my trades are standard or high conviction. Max conviction is rare and reserved for situations where all the data points are screaming and the narrative has clear momentum. I’ve learned that even when everything looks perfect, there’s still a meaningful chance of loss. So I never bet more than I can comfortably lose.
What This Workflow Looks Like in Practice
On a typical trading day, I have three dedicated screening sessions:
Morning (9-10AM EST): Check what happened overnight. Look for tokens that built during the quiet hours and are starting to get attention as traders wake up. This is often my most productive session.
Afternoon (1-2PM EST): Quick scan for anything new. The lunchtime period is when a lot of European and US activity overlaps, so there’s decent flow. Shorter session, usually 15-20 minutes.
Evening (8-9PM EST): Check how the day’s tokens are doing. Decide what to hold overnight and what to sell. Also scan for fresh launches that might build during the night.
Total active screening time: about 60-90 minutes per day. The rest of the time I’m not looking at my tracker. This structure prevents the “always on” burnout that kills so many memecoin traders.
Common Mistakes I Made Building This Workflow
This process didn’t come together overnight. I made a lot of errors refining it, and some of them might save you time.
Setting filters too loose. When I first started, I was afraid of missing winners, so my filters barely eliminated anything. I was overwhelmed with options and made worse decisions because I couldn’t evaluate them all properly. Tighter filters are almost always better than looser ones.
Skipping the social check. For a while, I relied entirely on on-chain data and ignored social presence. I kept buying tokens with great metrics but zero community, and they’d die within hours because nobody was talking about them. On-chain data tells you where a token is. Social activity tells you where it’s going.
Trading every day. Some days, nothing good shows up. For the first few months, I’d force trades on slow days because I felt like I should be doing something. Those forced trades had a terrible win rate. Now I’m comfortable with zero-trade days. The market isn’t going anywhere.
Not logging results. I didn’t start tracking my trades in a spreadsheet until month three. Once I did, I could see exactly which filter settings and screening criteria were producing winners. Data-driven refinement is how this workflow got to where it is. Without tracking, you’re just guessing at what works.
Build Your Own Version
My specific filter settings and timing work for me. They might not work for you, because your risk tolerance, available capital, time zone, and schedule are different. That’s fine.
What matters is the structure: automatic filters to cut the noise, a visual scan to find interesting candidates, a detail check to validate them, and consistent position sizing to manage risk. The specific parameters within each layer are personal and should evolve as you gain experience.
Start with the basics on a memecoin tracker, log your results, and adjust. The workflow that makes you money will reveal itself through the data if you’re paying attention.