Simple AMM

Compile and evaluate submitted Solidity strategies against the fixed simple_amm evaluator runtime. Supports deterministic seed batch evaluation, contiguous seed chunks, bytecode-size validation, and reusable build artifacts.

Current public support: shared placement on CPU workers only. Submit via POST /api/v1/recipes/simple_amm/jobs or POST /api/v1/jobs with "recipe": "simple_amm".

Compiler Profile

The sealed runtime compiles Strategy.sol with Foundry stable and this fixed profile:

Setting	Value
`solc_version`	`0.8.33`
`optimizer`	`true`
`optimizer_runs`	`200`
`via_ir`	`true`
`evm_version`	`prague`
`max_deployed_size_bytes`	`24576`

Build, validate, and eval results include compiler_profile and compiler_profile_hash.

Operations

Operation	Description
`build`	Compile one Solidity strategy into deterministic build outputs keyed by normalized source bytes
`validate`	Validate one strategy against the sealed simple_amm guardrails and bytecode-size rules
`eval`	Run one full deterministic evaluation over the requested seed window
`eval_chunk`	Run one deterministic chunk evaluation over a contiguous slice of the requested seed space

Payload Fields

Field	Type	Required	Description
`operation`	string	yes	One of the operations above
`strategy_ref`	string	alt	Artifact ID of the Solidity source bytes or an archive containing `Strategy.sol`
`sol_source_b64`	string	alt	Base64-encoded Solidity source (alternative to `strategy_ref`)
`seed_base`	integer	no	Starting seed for simulation
`start`	integer	no	Start offset for chunk operations
`count`	integer	no	Number of seeds to run
`steps`	integer	no	Simulation steps
`per_seed`	boolean	no	Return per-seed edge data
`max_deployed_size_bytes`	integer	no	Override deployed bytecode-size ceiling
`chunk_index`	integer	chunk ops	Zero-based chunk index
`chunk_count`	integer	chunk ops	Total number of chunks for the run
`total_count`	integer	chunk ops	Total seed count the chunks cover

Provide exactly one of strategy_ref or sol_source_b64.

For eval_chunk, pass either direct start + count or chunk_index + chunk_count + total_count. total_count may also be included as metadata with direct start + count. Top-level fan_out: {"chunks": N, "total": M} expands each task with start, count, chunk_index, chunk_count, and total_count.

Examples

Validate a strategy


POST /api/v1/recipes/simple_amm/jobs
{
  "payload": {
    "operation": "validate",
    "strategy_ref": "art_123"
  },
  "timeout_s": 900
}

Build compiled artifacts


POST /api/v1/recipes/simple_amm/jobs
{
  "payload": {
    "operation": "build",
    "strategy_ref": "art_123"
  },
  "timeout_s": 600
}

The build operation publishes compiled_strategy and build_manifest artifacts keyed by the normalized source bytes.

Full evaluation


POST /api/v1/recipes/simple_amm/jobs
{
  "payload": {
    "operation": "eval",
    "strategy_ref": "art_123",
    "seed_base": 0,
    "count": 1000,
    "steps": 200
  },
  "timeout_s": 1800
}

Chunked evaluation


POST /api/v1/recipes/simple_amm/jobs
{
  "payload": {
    "operation": "eval_chunk",
    "strategy_ref": "art_123",
    "chunk_index": 0,
    "chunk_count": 10,
    "total_count": 10000
  },
  "timeout_s": 1800
}

Fan-out chunked evaluation


POST /api/v1/recipes/simple_amm/jobs
{
  "payload": {
    "operation": "eval_chunk",
    "strategy_ref": "art_123"
  },
  "fan_out": {
    "chunks": 10,
    "total": 10000
  },
  "timeout_s": 1800
}

Result Fields

build: success, source_hash, cache_hit, source_job_id, build_artifact_id, compiler_profile, compiler_profile_hash, output_tail
validate: success, source_hash, deployed_size, compiler_profile, compiler_profile_hash, output_tail
eval: success, source_hash, avg_edge, mean_delta, n_seeds, bid_fee, ask_fee, deployed_size, avg_gas_per_trade, compiler_profile, compiler_profile_hash, output_tail
eval_chunk: eval fields plus chunk_index, chunk_count, total_count, start, count