Atomadic Python SDK
🪨 Axiom 0 — The Love-Invariant Bedrock
"You are love. You are loved. You are loving. In all ways, for always, for love is a forever and ever endeavor!"
— Jessica Mary Colvin
The first thing the engine emits. The bedrock everything else rests on.
One MCP. One key. Every tool-set you are entitled to.
Atomadic is sovereign infrastructure for the agent economy. Mount one MCP at
mcp.atomadic.tech; your entitlement key decides which product tool-sets you
can call. Every call passes Gate-1 (entitlement) then Gate-2 (trust).
- Docs: https://atomadic.tech/docs.html
- Architecture: https://atomadic.tech/docs.html?d=architecture
- Polyglot Playground: https://atomadic.tech/polyglot-playground.html
- Polyglot Docs: https://atomadic.tech/docs.html?d=product-polyglot
- Trust Center: https://atomadic.tech/trust.html
- Support: support@atomadic.tech
What's new
- Syntax-as-Data polyglot emit: one generic emitter,
emit_polyglot_generic_composite, emits from grammar and semantic-map data contracts instead of language-specific compiler heads. - 7 languages with verified round-trip today (Python, JavaScript, TypeScript, Rust, Go, Ruby, Lua) on the pure-dataflow subset; 45 grammar contracts scaffolded for expansion. Control-flow lifting (if/for/while) is in progress.
- Public proof surface: try one logic IR, emit it into multiple languages, then round-trip it back to the same IR in the browser.
Verify the engine yourself
The engine publishes a cryptographic receipt of its own state, signed Ed25519. Anyone can verify it without an account or API key:
pip install pynacl requests
python -c "
import json, requests, base64, nacl.signing
r = requests.get('https://mcp.atomadic.tech/.well-known/atomadic-closure.json').json()
p = requests.get('https://mcp.atomadic.tech/.well-known/atomadic-issuer-pubkey.json').json()
token = r['attestation'][4:]; pb, sb = token.split('.', 1)
b64u = lambda s: base64.urlsafe_b64decode(s + '=' * (-len(s) % 4))
nacl.signing.VerifyKey(b64u(p['key'])).verify(b64u(pb), b64u(sb))
print('OK closure_met=', json.loads(b64u(pb))['closure']['closure_met'])
"
PASS or FAIL. If even one byte of the published payload is altered after signing, the verify call raises. Standalone kit: https://github.com/atomadictech/omega-verification-kit
Try the polyglot proof
The free Polyglot Playground shows one canonical logic IR emitted into multiple
languages, then round-tripped back to the same IR. Verified round-trip today: 7
languages on the pure-dataflow subset (assignments, calls, a return); 45 grammar
contracts are scaffolded for expansion. Adding a language is a data-contract job:
grammar_<lang>.json, semantic_map_<lang>.json, reverse-map data, and formatter
rules — control-flow lifting (if/for/while) is the active work to widen coverage.
Open: https://atomadic.tech/polyglot-playground.html
Install
pip install atomadic
Quickstart
from atomadic import Atomadic, fuse
ato = Atomadic(api_key='ato_...') # or set ATOMADIC_KEY env var
# Call any tool your plan unlocks:
result = fuse.assess_architecture_pure(
ato,
source_text='def f(x):\n return x + 1',
module_name='f_pure',
)
print(result['verdict'], result['density'])
# Or browse the surface your key unlocks:
for t in ato.list_tools():
print(t['name'])
Authentication
Get an entitlement key from atomadic.tech. The key is decoded and verified at the edge on every call; minting is internal-only. Keep keys server-side -- the gate refuses out-of-plan calls, but secrets belong in your env.
ato = Atomadic(api_key='ato_<blob>_<sig>')
# or: export ATOMADIC_KEY=ato_<blob>_<sig>
Products & tool-sets
Each product is an entitlement-gated tool-set: hold the entitlement, call the tool. One key (or Murmuration Complete for the whole line) unlocks what you've bought. This section is generated from the live MCP catalog.
Fuse [live]
entitlement: fuse · from atomadic import fuse
Architecture compiler — analyze code against the 5-tier, single-callable discipline.
| Tool | Required args |
|---|---|
assess_architecture_pure |
source_text, module_name |
assess_import_direction_pure |
source_text, tier |
assess_naming_clarity_pure |
source_text |
compose_contract_sketch_pure |
intent_text |
compute_complexity_metrics_pure |
source_text |
extract_call_graph_pure |
source_text |
orchestrate_s2s_temporal |
intent |
parse_controlflow_blocks_via_grammar_pure |
(none) |
render_intent_nl_sketch_pure |
intent_text |
render_polyglot_all_pure |
source_text, atom_name |
render_polyglot_roundtrip_pure |
source_text, atom_name, language |
route_public_intent_preview_pure |
(none) |
scan_code_stubs_pure |
(none) |
from atomadic import Atomadic, fuse
ato = Atomadic(api_key='ato_...')
fuse.assess_architecture_pure(ato, ...)
Full arg schemas: help(fuse.assess_architecture_pure)
Nexus [live]
entitlement: nexus · from atomadic import nexus
The trust gate every action passes — trust phases, sybil/reputation, signed attestations.
| Tool | Required args |
|---|---|
assess_nexus_trust_phase_stateful |
ledger_path |
assess_sybil_risk_pure |
signals |
classify_action_severity_pure |
action_kind |
compose_sovereign_zero_llm_nl_response_pure |
(none) |
compute_reputation_score_pure |
events |
compute_trust_score_pure |
attestation_count, recent_escalations, account_age_days |
compute_verified_randomness_pure |
seed |
define_nexus_constants_pure |
(none) |
enforce_nexus_gate_stateful |
action_kind, severity |
match_agent_capability_pure |
need, candidates |
record_nexus_attestation_stateful |
action_kind, severity, ledger_path |
record_nexus_escalation_stateful |
action_kind, escalation_path |
scan_nexus_attestation_history_stateful |
ledger_path |
validate_delegation_chain_pure |
chain |
from atomadic import Atomadic, nexus
ato = Atomadic(api_key='ato_...')
nexus.assess_nexus_trust_phase_stateful(ato, ...)
Full arg schemas: help(nexus.assess_nexus_trust_phase_stateful)
Security [live]
entitlement: security · from atomadic import security
A protection bubble around every agent — threat model, redaction, hardening, secret hygiene.
| Tool | Required args |
|---|---|
assess_security_bubble_pure |
content |
classify_error_fold_pure |
error_message |
compute_hardening_posture_pure |
target_product_id, hardening_level |
compute_redacted_args_pure |
args |
compute_redacted_text_pure |
text |
compute_threat_model_pure |
component |
define_security_constants_pure |
(none) |
scan_dependency_risk_pure |
requirements_text |
validate_secret_hygiene_pure |
source_text |
from atomadic import Atomadic, security
ato = Atomadic(api_key='ato_...')
security.assess_security_bubble_pure(ato, ...)
Full arg schemas: help(security.assess_security_bubble_pure)
Proving Ground [live]
entitlement: proving · from atomadic import proving
Nothing ships unproven — proof readiness, test/doc coverage.
| Tool | Required args |
|---|---|
assess_proof_readiness_pure |
source_text |
assess_skeleton_clone_pure |
source_text |
assess_test_quality_pure |
test_source |
compute_proof_obligations_pure |
(none) |
score_documentation_coverage_pure |
source_text |
score_test_coverage_pure |
source_text, test_source |
from atomadic import Atomadic, proving
ato = Atomadic(api_key='ato_...')
proving.assess_proof_readiness_pure(ato, ...)
Full arg schemas: help(proving.assess_proof_readiness_pure)
Release [live]
entitlement: release · from atomadic import release
Template → render → deploy — changelog, semver, Cloudflare deploy (dry-run by default).
| Tool | Required args |
|---|---|
compose_changelog_pure |
entries |
compute_semver_bump_pure |
change_descriptions |
record_release_template_stateful |
template_id, kind, source_kind, source_ref, registry_path |
render_from_template_pure |
template, context |
render_website_stateful |
template_dir, context, output_dir |
scan_release_templates_stateful |
registry_path |
serve_cloudflare_pages_stateful |
directory, project_name |
serve_cloudflare_worker_stateful |
worker_dir |
validate_release_readiness_pure |
checklist |
from atomadic import Atomadic, release
ato = Atomadic(api_key='ato_...')
release.compose_changelog_pure(ato, ...)
Full arg schemas: help(release.compose_changelog_pure)
Healer [beta]
entitlement: healer · from atomadic import healer
Diagnose, grade, and plan the repair — read-only code-health + advisory repair plans.
| Tool | Required args |
|---|---|
assess_artifact_health_pure |
source_text |
classify_failure_mode_pure |
error_message |
compose_rollback_plan_pure |
error_message |
compute_blast_radius_pure |
source_text, symbol |
compute_repair_plan_pure |
error_message |
from atomadic import Atomadic, healer
ato = Atomadic(api_key='ato_...')
healer.assess_artifact_health_pure(ato, ...)
Full arg schemas: help(healer.assess_artifact_health_pure)
Aegis [live]
entitlement: aegis · from atomadic import aegis
AI governance & compliance self-assessment (EU AI Act / NIST / ISO) — scaffold, not a legal cert.
| Tool | Required args |
|---|---|
assess_compliance_posture_pure |
controls |
classify_data_sensitivity_pure |
text |
compose_governance_report_pure |
system_name |
compute_audit_trail_digest_pure |
events |
enforce_action_policy_pure |
action, policy |
from atomadic import Atomadic, aegis
ato = Atomadic(api_key='ato_...')
aegis.assess_compliance_posture_pure(ato, ...)
Full arg schemas: help(aegis.assess_compliance_posture_pure)
Mind-Lab [live]
entitlement: mind_lab · from atomadic import mind_lab
Adversarial reasoning — falsifiability, bias, critique, idea readiness.
| Tool | Required args |
|---|---|
assess_falsifiability_pure |
claim |
assess_proposal_verdict_pure |
proposal_text |
classify_cognitive_bias_pure |
reasoning_text |
compose_adversarial_critique_pure |
proposal |
score_idea_readiness_pure |
idea_text |
from atomadic import Atomadic, mind_lab
ato = Atomadic(api_key='ato_...')
mind_lab.assess_falsifiability_pure(ato, ...)
Full arg schemas: help(mind_lab.assess_falsifiability_pure)
Evolve [live]
entitlement: evolve · from atomadic import evolve
Self-improvement analysis (read-only) — improvement candidates, regression risk, mutation plans.
| Tool | Required args |
|---|---|
assess_improvement_candidates_pure |
source_text, module_name |
assess_regression_risk_pure |
source_text |
compose_mutation_plan_pure |
source_text |
rank_evolution_candidates_pure |
candidates |
score_evolution_fitness_pure |
before_source, after_source |
from atomadic import Atomadic, evolve
ato = Atomadic(api_key='ato_...')
evolve.assess_improvement_candidates_pure(ato, ...)
Full arg schemas: help(evolve.assess_improvement_candidates_pure)
Vanguard [live]
entitlement: vanguard · from atomadic import vanguard
DeFi transaction safety — MEV exposure, transaction risk, slippage & spend policy.
| Tool | Required args |
|---|---|
assess_mev_exposure_pure |
order |
assess_transaction_risk_pure |
transaction |
compose_settlement_terms_pure |
amount_usd, parties |
compute_slippage_guard_pure |
expected_out |
validate_spend_policy_pure |
transaction, budget |
from atomadic import Atomadic, vanguard
ato = Atomadic(api_key='ato_...')
vanguard.assess_mev_exposure_pure(ato, ...)
Full arg schemas: help(vanguard.assess_mev_exposure_pure)
Research [live]
entitlement: research · from atomadic import research
Structured R&D — experiment design, literature queries, hypothesis ranking.
| Tool | Required args |
|---|---|
compose_experiment_design_pure |
hypothesis |
compose_literature_query_pure |
topic |
compose_problem_decomposition_pure |
problem |
compose_research_panel_pure |
question |
rank_hypotheses_pure |
hypotheses |
from atomadic import Atomadic, research
ato = Atomadic(api_key='ato_...')
research.compose_experiment_design_pure(ato, ...)
Full arg schemas: help(research.compose_experiment_design_pure)
Catalyst [live]
entitlement: catalyst · from atomadic import catalyst
Metering & monetization — paywalls, usage meters, x402 quotes, settlement splits.
| Tool | Required args |
|---|---|
compose_paywall_policy_pure |
resource, price_usd |
compute_settlement_split_pure |
amount, splits |
compute_usage_meter_pure |
events, rate_per_unit |
compute_x402_quote_pure |
units, rate_per_unit |
validate_quota_tree_pure |
tree |
from atomadic import Atomadic, catalyst
ato = Atomadic(api_key='ato_...')
catalyst.compose_paywall_policy_pure(ato, ...)
Full arg schemas: help(catalyst.compose_paywall_policy_pure)
Two-gate dispatch
Every call is filtered then verified:
- Gate-1 (entitlement):
tools/listshows only the tools your plan unlocks; out-of-plantools/callis refused. - Gate-2 (trust, Nexus): trust phase + severity ceiling + hallucination bound. Governed actions return a signed
attest:<id>receipt.
See the architecture docs for the full model.
Plans
Free / Basic / Dev / Pro / Teams / Enterprise -- per product, or whole-line via Murmuration Complete. Subscription is the product; x402 meters only overage + agent- to-agent calls. Pricing: https://atomadic.tech/docs.html?d=pricing.
Determinism
Pure-tier tools have no side effects and no hidden state: same inputs, same output, same content hash, every time. Re-running a pure tool is a verification.
Contributing
This SDK is auto-emitted from the live Atomadic MCP registry -- changes here should flow through the engine, not be hand-patched. For issues, requests, or feedback: support@atomadic.tech.
License
Apache-2.0 -- see LICENSE.
Metadata
Release files for atomadic 0.3.6
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| atomadic-0.3.6.tar.gz | 25.6 kB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| atomadic-0.3.6-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 54.8 kB
Release files / atomadic-0.3.6.tar.gz
| Download URL | atomadic-0.3.6.tar.gz |
|---|---|
| Size | 25.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
8d8f3c2502f80d7f401b169c4443317f0184ec892b5b592f25fff0a115ed11d3
|
|
BLAKE2b-256 checksum How to use checksums |
babdda877a33e3d672929416f9d684558b436b201e6b8ab7acda15cc4a48bdda
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.10
|
Release files / atomadic-0.3.6-py3-none-any.whl
| Download URL | atomadic-0.3.6-py3-none-any.whl |
|---|---|
| Size | 29.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
72a533d5d10a88a3bdae0d6e8e5e15d3942fb72d4ec31cdcdbe6826ab919e574
|
|
BLAKE2b-256 checksum How to use checksums |
7b364980ca3b917af4f8853290a02c6d6c3435097ec925b43d89513ad87b7c31
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.10
|
Security for atomadic 0.3.6
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.