# Lazareus

> Network health and priority fee data for Solana, Base, and Ethereum,
> built for AI agents. Every endpoint answers a question in a few hundred
> milliseconds — no dashboard, no account, no subscription. Pay per call
> via x402, on Solana or Base (whichever wallet you have) — independent
> from which chain's data you're asking about.

Lazareus is not a dashboard. It's an API designed to be called by AI agents
before they act — before executing a timing-sensitive transaction, before
attaching a priority fee. Responses are small, fast, and structured for
machine consumption first (a `format=llm` mode is also available for a
natural-language sentence instead of JSON).

## Access

- Human docs: https://lazareus.xyz/docs.html
- MCP server (Streamable HTTP): https://lazareus.xyz/mcp
- REST base URL: https://lazareus.xyz

## Payment

- Protocol: x402 (HTTP 402 Payment Required)
- Networks: Solana (mainnet) or Base (mainnet) — pick whichever you have a wallet for
- Currency: USDC
- Price: $0.002 per call, no account or subscription required
- On an unpaid request, the API responds `402` with a `PAYMENT-REQUIRED`
  header describing the exact payment terms (amount, asset, payTo,
  network). Sign and retry with a `PAYMENT-SIGNATURE` header to get the
  real response.

## Endpoints

### Solana

### GET /get_network_health/solana

Returns a 0-100 Solana network health score combining average slot time,
% of delinquent stake, median prioritization fee (congestion signal), and
skipped-slot rate (block production health), plus the raw metrics and
per-metric sub-scores behind it. Useful before executing a
timing-sensitive transaction.

Query parameters:
- `format=llm` (optional) — returns a natural-language sentence instead of JSON

Example:
```
GET https://lazareus.xyz/get_network_health/solana
```

Response:
```json
{
  "score": 94,
  "status": "healthy",
  "degraded": false,
  "reasons": [],
  "metrics": {
    "avg_slot_time_ms": 421.2,
    "delinquent_stake_pct": 0.06,
    "median_prioritization_fee_microlamports": 0,
    "skip_rate_pct": 0
  },
  "sub_scores": {
    "slot_time": 89,
    "delinquent_stake": 99,
    "prioritization_fee": 100,
    "skip_rate": 100
  },
  "thresholds": {
    "slot_time_degraded_ms": 600,
    "delinquent_stake_degraded_pct": 10,
    "prioritization_fee_degraded_microlamports": 10000,
    "skip_rate_degraded_pct": 5
  },
  "checked_at": "2026-08-06T18:45:34.680Z"
}
```

### GET /get_recommended_priority_fee/solana

Recommends a Solana transaction priority fee (in micro-lamports per
compute unit) based on recent network activity, at low/medium/high
confidence levels. Multiply the returned value by your transaction's
compute unit budget to get the actual priority fee to attach.

Query parameters:
- `format=llm` (optional) — returns a natural-language sentence instead of JSON

Example:
```
GET https://lazareus.xyz/get_recommended_priority_fee/solana
```

Response:
```json
{
  "recommended_microlamports_per_cu": 150,
  "levels": { "low": 50, "medium": 150, "high": 400 },
  "sample_size": 20,
  "checked_at": "2026-08-06T18:45:34.680Z"
}
```

### Base & Ethereum (EVM)

Same signal, computed from EIP-1559 chain data instead of Solana's slot/
stake/prioritization-fee metrics: average block time, average gas-used
ratio (how full recent blocks are — a congestion signal), base fee
volatility (coefficient of variation of recent baseFeePerGas samples),
and median priority fee in gwei.

### GET /get_network_health/base
### GET /get_network_health/ethereum

Returns a 0-100 network health score for Base or Ethereum mainnet
respectively, plus the raw metrics and per-metric sub-scores behind it.
Same query parameters and response shape as `/get_network_health/solana`,
with an added `"chain"` field (`"base"` or `"ethereum"`) and EVM-specific
`metrics`/`thresholds` keys (`avg_block_time_ms`, `avg_gas_used_ratio`,
`base_fee_volatility`, `median_priority_fee_gwei`).

Example:
```
GET https://lazareus.xyz/get_network_health/base
```

Response:
```json
{
  "chain": "base",
  "score": 96,
  "status": "healthy",
  "degraded": false,
  "reasons": [],
  "metrics": {
    "avg_block_time_ms": 2010.4,
    "avg_gas_used_ratio": 0.42,
    "base_fee_volatility": 0.08,
    "median_priority_fee_gwei": 0.004
  },
  "sub_scores": {
    "block_time": 99,
    "gas_used_ratio": 100,
    "base_fee_volatility": 84,
    "priority_fee": 100
  },
  "thresholds": {
    "block_time_degraded_ms": 4000,
    "gas_used_ratio_degraded": 0.95,
    "base_fee_volatility_degraded": 0.5,
    "priority_fee_degraded_gwei": 2
  },
  "checked_at": "2026-08-14T18:00:00.000Z"
}
```

### GET /get_recommended_priority_fee/base
### GET /get_recommended_priority_fee/ethereum

Recommends an EIP-1559 priority fee (`maxPriorityFeePerGas`, in gwei) for
Base or Ethereum mainnet respectively, at low/medium/high confidence
levels, plus the current base fee.

Example:
```
GET https://lazareus.xyz/get_recommended_priority_fee/base
```

Response:
```json
{
  "chain": "base",
  "current_base_fee_gwei": 0.0132,
  "recommended_priority_fee_gwei": 0.0041,
  "levels": { "low": 0.001, "medium": 0.0041, "high": 0.0098 },
  "sample_size": 20,
  "checked_at": "2026-08-14T18:00:00.000Z"
}
```

## Errors

All endpoints share a standard error shape:
```json
{ "error": { "code": "internal_error", "message": "..." } }
```

## Free, unpaid endpoints

- `GET /health` — liveness check
- `GET /stats` — total calls served, average processing time (excludes x402 payment settlement)
- `GET /uptime` — hourly uptime history, last 30 days, per chain (`network`, `network_base`, `network_ethereum`)

## MCP tools

The MCP server at `https://lazareus.xyz/mcp` (Streamable HTTP transport)
exposes the same six checks as tools:

- `get_network_health()`
- `get_recommended_priority_fee()`
- `get_network_health_base()`
- `get_recommended_priority_fee_base()`
- `get_network_health_ethereum()`
- `get_recommended_priority_fee_ethereum()`

Same underlying logic and response shape as the REST endpoints above.