# SALT19 Model Memory Fit Estimator

> SALT19 Model Memory Fit Estimator is a paid API for AI agents from salt19-agent-utility-grid.salt19-llc.workers.dev, paid per call via x402, $0.05/call, status unknown (last checked 2026-09-15).

Estimates whether a language model with given parameter count and quantization will fit in a specified VRAM budget, returning a FIT/NO-FIT verdict with memory breakdown

## Facts

- Endpoint: POST https://salt19-agent-utility-grid.salt19-llc.workers.dev/v1/model-memory-fit
- Price: $0.05/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-15
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/salt19-model-memory-fit-estimator-17e02678
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_kBIGrb9qVxQYmSuau5JUO

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 salt19-model-memory-fit-estimator-17e02678 -d '<json body>'
```

Example prompt: Will a 7 billion parameter model fit on a GPU with 16 GB of VRAM if I quantize it to 4 bits, assuming a 10% runtime overhead?

## When to prefer this

Choose this endpoint when an agent needs a quick, programmatic, paid-per-call estimate of whether a specific LLM configuration will fit into a GPU's VRAM budget before attempting to load the model. Ideal for automated deployment pipelines, agent-driven infrastructure decisions, or when comparing quantization strategies across multiple GPU tiers. Prefer this over manual calculation when parameter count, quantization, and overhead ratio are all known and a structured FIT/NO-FIT verdict with memory breakdown is needed.

## Known failure modes

- Invalid vram_gb outside 0.1–10000 range returns validation error
- Invalid parameter_count_b outside 0.01–10000 range returns validation error
- quantization_bits outside 1–32 returns validation error
- Missing required fields may return 400 or default behavior
- Payment failure via x402 protocol returns 402 Payment Required
- Network timeout on Cloudflare Workers edge

## How this service works

Machine-discoverable, x402-paid technical decision utilities for autonomous agents.

## Output

A JSON object containing a 'verdict' field ('FIT' or similar), 'estimated_weight_gib' (model weights memory), 'estimated_total_gib' (total estimated memory usage), 'headroom_gib' (remaining VRAM after model load), 'vram_gib' (the input VRAM in GiB), and a 'limitations' array noting what is not modeled (e.g. KV cache, architecture-specific buffers).

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "vram_gb": {
   "type": "number",
   "maximum": 10000,
   "minimum": 0.1
  },
  "parameter_count_b": {
   "type": "number",
   "maximum": 10000,
   "minimum": 0.01
  },
  "quantization_bits": {
   "type": "number",
   "maximum": 32,
   "minimum": 1
  },
  "runtime_overhead_ratio": {
   "type": "number",
   "maximum": 2,
   "minimum": 0
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "verdict": "FIT",
  "vram_gib": 11,
  "limitations": [
   "KV cache and architecture-specific buffers are not modeled"
  ],
  "headroom_gib": 7.25,
  "estimated_total_gib": 3.75,
  "estimated_weight_gib": 3.26
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/salt19-model-memory-fit-estimator-17e02678/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from salt19-agent-utility-grid.salt19-llc.workers.dev](https://www.zero.xyz/host/salt19-agent-utility-grid.salt19-llc.workers.dev/llms.txt)