SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.19633 · 2024

OptiMUS-0.3: Using Large Language Models to Model and Solve Optimization Problems at Scale

arXiv · PDF · Open in the Atlas

Code that ran

14 of 15 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 13 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2407.19633.json · badge: /api/ran/2407.19633.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
teshnizi/optimus canonical (classifier label)
classifier basis: census
14 ran · 1 no recorded run
FunctionStatusWhere it lives
execute_code Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/Reflexion.py
code served (permissive licence) · get_code("0a6804b25b53643a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import subprocess
import tempfile

# Create a temporary Python file that will be executed
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, 'test_script.py')

# Write a simple Python script that outputs something
with open(temp_file, 'w') as f:
    f.write('print("100000")')

ARGS = {
    'file_path': temp_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
execute_code Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/execute_code.py
code served (permissive licence) · get_code("dab23c4f1db56c16")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
import subprocess

# Create a temporary directory for the test
temp_dir = tempfile.mkdtemp()

# Create a simple Python script to execute
code_content = """
print("42000")
"""

code_filename = "test_script.py"
code_path = os.path.join(temp_dir, code_filename)

with open(code_path, 'w') as f:
    f.write(code_content)

ARGS = {
    'dir': temp_dir,
    'code_filename': code_filename
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_code Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/Reflexion.py
code served (permissive licence) · get_code("3400042ef10a04a1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

def extract_code(text):
    ind1 = text.find('=====')
    ind2 = text.find('=====', ind1 + 5)
    code = text[ind1 + 5:ind2].strip()
    code = code.replace('```python', '').replace('```', '').strip()
    return code

# Construct test input
sample_text = """
Some documentation here.
=====

def hello():
    return "world"

=====
More text after.
"""

ARGS = {
    'text': sample_text
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_code Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/execute_code.py
code served (permissive licence) · get_code("1b35aeee4688c891")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

# Analyze the function: extract_code takes a 'text' parameter (string)
# and uses string methods: find(), slicing, strip(), replace()
# So we need to construct a valid text string that contains the delimiters

text = """
Some preamble content here.
=====

def hello_world():
    print("Hello, World!")

=====
Some conclusion content here.
"""

ARGS = {
    "text": text
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_code_from_end Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/target_code.py
code served (permissive licence) · get_code("e55cc7e0d69fd4cc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

# Analyze the function: extract_code_from_end(text)
# The function takes a single parameter 'text' which is a string
# It searches for code blocks delimited by either '=====' or '```python' markers
# and extracts the code between them, with various cleanup operations

# Test case 1: text with ===== delimiters
text_with_equals = """
Some preamble text here
===== 
def hello():
    print("world")
=====
Some text after
"""

# Test case 2: text with ```python delimiters
text_with_backticks = """
Some preamble text here

def hello():
    print("world")

Some text after
"""

# Test case 3: text with nested markdown-style code
text_with_nested = """
Some preamble

def foo():
    # some code
    ```python
    nested = True
    

"""

ARGS = {
    "text": text_with_equals
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_json_from_end Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from Forrest-Stone/SAC-Opt, is pointer only (licence: NONE) · get_code("1b919c0f41cadd94")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Forrest-Stone/SAC-Opt, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_json_from_end_backup Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from Forrest-Stone/SAC-Opt, is pointer only (licence: NONE) · get_code("8db12a3fbdb85d30")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Forrest-Stone/SAC-Opt, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_keep_remove Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/variables.py
code served (permissive licence) · get_code("637439f058a00607")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

# Analyze the function:
# - text: string that is checked with .strip() and .endswith()
# - vars: dictionary that is accessed with del vars[var] and returned
# - var: string used in format() call and as dict key

# Setup valid arguments
text = "  some configuration KEEP  "
vars_dict = {
    "model_size": 128,
    "learning_rate": 0.001,
    "batch_size": 32
}
var = "model_size"

ARGS = {
    "text": text,
    "vars": vars_dict,
    "var": var
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_list_from_end Ran
ran, honoured the contract we drafted
teshnizi/optimus/utils.py
pointer only (licence: MIT); the stored copy get_code serves, from Forrest-Stone/SAC-Opt, is pointer only (licence: NONE) · get_code("dfe55bee535745f0")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from Forrest-Stone/SAC-Opt, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

extract_objective Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/objective.py
code served (permissive licence) · get_code("2a3974ba00dea130")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

def extract_objective(text):
    ind_1 = text.find('=====')
    ind_2 = text.find('=====', ind_1 + 1)
    objective = text[ind_1:ind_2]
    objective = objective.replace('=====', '').strip()
    objective = objective.replace('OBJECTIVE:', '').strip()
    return objective

# Construct a valid text input with the expected format
text = "Some preamble here\n===== OBJECTIVE: Maximize accuracy and minimize loss ===== More text after"

ARGS = {
    'text': text
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_score Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/parameters.py
code served (permissive licence) · get_code("edbd5ba229bb9dfc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import re

torch.manual_seed(0)

# Analyzing the function extract_score:
# - text: a string that will be searched for pattern '\d out of 5'
# - params: a dict that may be modified (del params[param])
# - param: a string key that identifies which parameter in params dict

# Create test data
text = "The model performance is 4 out of 5 stars."
params = {
    "learning_rate": 0.001,
    "batch_size": 32,
    "hidden_dim": 128
}
param = "learning_rate"

ARGS = {
    "text": text,
    "params": params,
    "param": param
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_score Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/variables.py
code served (permissive licence) · get_code("90629c8ff26c53f1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

def extract_score(text, vars, var):
    match = re.search('\\d out of 5', text.lower())
    if match:
        score = int(match.group()[0])
        if score > 3:
            return (True, vars)
        else:
            inp = input('LLMs reasoning: {}\n Do you want to keep variable {}? (y/n): '.format(text, var))
            if inp == 'y':
                return (True, vars)
            else:
                del vars[var]
                return (True, vars)
    else:
        return (False, None)


# Setup code to construct valid arguments
text = "The model performed well, scoring 4 out of 5 on the benchmark."
vars = {"model_v1": "checkpoint_1", "model_v2": "checkpoint_2"}
var = "model_v1"

ARGS = {
    "text": text,
    "vars": vars,
    "var": var
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_param_code Ran
ran; our contract draft was wrong, not the code
teshnizi/optimus/generate_code.py
code served (permissive licence) · get_code("729760725b6df4b3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Analyzing the function:
# get_param_code takes three parameters: symbol (str), shape (any), definition (str)
# The function simply formats them into an f-string
# No complex objects or method calls are needed

ARGS = {
    "symbol": "weight",
    "shape": (4, 3),
    "definition": "parameter matrix for linear layer"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_var_code Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
teshnizi/optimus/generate_code.py
code served (permissive licence) · get_code("a6bbc053b1dfeae4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - symbol: a string used as variable name
# - shape: a list (can be empty or contain integers)
# - type: a string representing variable type (will be uppercased and used in GRB.{type})
# - definition: a parameter that appears in the function signature but is not used in the body
# - solver: a string, defaults to 'gurobipy'

ARGS = {
    'symbol': 'x',
    'shape': [3, 4],
    'type': 'continuous',
    'definition': None,
    'solver': 'gurobipy'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_score_constraint_model Not yet run
unverified (harvested, no recorded run)
teshnizi/optimus/constraint_model.py
code served (permissive licence) · get_code("4667ece67c9e29d5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 12 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
extract_json_from_end 1 other paper
1 later
2510.05115 (2025-10)
SAC-Opt: Semantic Anchors for Iterative Correction in Optimization Modeling
extract_json_from_end_backup 1 other paper
1 later
2510.05115 (2025-10)
SAC-Opt: Semantic Anchors for Iterative Correction in Optimization Modeling
extract_list_from_end 1 other paper
1 later
2510.05115 (2025-10)
SAC-Opt: Semantic Anchors for Iterative Correction in Optimization Modeling

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Optimization problems are pervasive in sectors from manufacturing and distribution to healthcare. However, most such problems are still solved heuristically by hand rather than optimally by state-of-the-art solvers because the expertise required to formulate and solve these problems limits the widespread adoption of optimization tools and techniques. We introduce a Large Language Model (LLM)-based system designed to formulate and solve (mixed integer) linear programming problems from their natural language descriptions. Our system can develop mathematical models, write and debug solver code, evaluate the generated solutions, and improve efficiency and correctness of its model and code based on these evaluations. OptiMUS is designed as a productivity tool for optimization practitioners who understand the problem domain and can describe it precisely, but seek to accelerate the modeling and implementation workflow. OptiMUS-0.3 utilizes a modular structure to process problems, allowing it to handle problems with long descriptions and complex data without long prompts. Experiments demonstrate that OptiMUS-0.3 outperforms direct-prompting baselines by over 43% on easy and 18% on hard instances. It remains competitive with fine-tuned specialist models on benchmark problems, and outperforms them on real-world case studies (28.6% vs. 0%) where fine-tuned models fail to generalize. Ablation studies show that modular architecture with error correction is central to these gains. A key finding is that system architecture is a stronger driver of performance than model capability. Structured decomposition with targeted error correction enables weaker models to match stronger models under naive prompting, and remains competitive with fine-tuned specialist models without retraining costs.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2407.19633")
get_code_for_paper("2407.19633")
have("2407.19633")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2407.19633.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2407.19633.svg)](https://syntology.ai/paper/2407.19633)

Connect an agent — have() is free.