# Avasis Solana Priority Fee Tips

> Avasis Solana Priority Fee Tips is a paid API for AI agents from agents.avasis.ai, paid per call via x402, $0.002/call, status unknown (last checked 2026-10-31).

Returns real-time Solana transaction priority-fee percentiles (p25/p50/p75/p90/max) and low/medium/high/veryHigh micro-lamport suggestions, refreshed every 2 seconds from the last ~150 slots

## Facts

- Endpoint: GET https://agents.avasis.ai/v1/tip/sol-priority-fee?utm_source=zero.xyz
- Price: $0.002/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-31
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/avasis-solana-priority-fee-tips-3e3ef724
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_wu1twaOo0x8fPtL7CWQhi

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability avasis-solana-priority-fee-tips-3e3ef724
```

Example prompt: What micro-lamport priority fee should I attach to my Solana transaction right now — give me the p75 percentile and the 'high' suggested tip level so I can set the compute unit price before submitting.

## When to prefer this

Use this endpoint when an AI agent is building or submitting Solana transactions and needs a real-time, percentile-based priority fee recommendation rather than a static default. It is preferable to raw RPC calls because it aggregates ~150 slots of data, provides tiered suggestions (low/medium/high/veryHigh), and supports efficient long-polling to detect fee changes. Choose this over generic Solana RPC if you need sub-second fee awareness without managing your own getRecentPrioritizationFees aggregation logic.

## Known failure modes

- Payment required (402) if prepaid meter balance is exhausted
- Stale data (stale:true) if Solana RPC is lagging or endpoint hasn't refreshed within 2s window
- Long-poll timeout if no new data arrives within the specified wait window
- Empty or zero percentiles (p25/p50 may be 0) during low-congestion periods, which is valid but may mislead fee selection
- Rate limiting or connection errors if polling too aggressively without using the long-poll mechanism

## How this service works

Solana priority-fee tip for agents that build transactions: p25/p50/p75/p90/max and non-zero percentiles of getRecentPrioritizationFees over the last ~150 slots, plus low/medium/high/veryHigh micro-lamport suggestions, refreshed every 2 s. Long-poll with ?since=<seq>&wait=25. Prepaid meter price $0.0002/call (packs at /v1/meter); SSE at /v1/stream/tips.

## Output

A JSON object containing: a sequence number (seq), the age of the data in milliseconds (ageMs), a staleness flag, the number of slot samples used (~150), priority fee percentiles (p25, p50, p75, p90, max) in micro-lamports per compute unit, and four tiered suggestions (low, medium, high, veryHigh) also in micro-lamports per compute unit. The observedAt timestamp and a changed flag indicate whether the data updated since the last poll.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "properties": {
   "type": "string"
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "seq": 812,
  "data": {
   "unit": "micro-lamports per compute unit",
   "samples": 150,
   "suggested": {
    "low": 1000,
    "high": 50000,
    "medium": 10000,
    "veryHigh": 200000
   },
   "percentiles": {
    "max": 2000000,
    "p25": 0,
    "p50": 0,
    "p75": 1000,
    "p90": 50000
   }
  },
  "ageMs": 420,
  "stale": false,
  "topic": "sol-priority-fee",
  "changed": true,
  "observedAt": "2026-10-29T12:00:00.000Z"
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/avasis-solana-priority-fee-tips-3e3ef724/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agents.avasis.ai](https://www.zero.xyz/host/agents.avasis.ai/llms.txt)