Skip to main content

agentscore-py

PyPI version License: MIT

The official Python client for AgentScore. Identity verification and compliance gating for agentic commerce (KYC, age, sanctions, jurisdiction), sync and async.

Install

pip install agentscore-py

Quick Start

from agentscore import AgentScore

client = AgentScore(api_key="as_live_...")

# Look up cached reputation (free)
rep = client.get_reputation("0x1234...")
print(rep["score"]["value"], rep["score"]["grade"])

# Filter to a specific chain
base_rep = client.get_reputation("0x1234...", chain="base")

# Identity gate with policy (paid)
gated = client.assess(
    "0x1234...",
    policy={
        "require_kyc": True,
        "require_sanctions_clear": True,
        "min_age": 21,
    },
)

if gated["decision"] == "deny":
    print(gated["decision_reasons"])  # ["kyc_required"]
    print(gated.get("verify_url"))  # URL for operator verification

# Check verification level
rep = client.get_reputation("0x1234...")
print(rep.get("verification_level"))  # "none" | "wallet_claimed" | "kyc_verified"

Credential-Based Identity

Agents without wallets can use operator credentials for identity:

result = client.assess(operator_token="opc_...")
print(result["decision"])  # "allow" | "deny"

Verification Sessions

Bootstrap identity for first-time agents. The success body carries structured next_steps (with action: "deliver_verify_url_and_poll") and a cross-merchant agent_memory hint. Poll responses carry next_steps.action from the typed NextStepsAction Literal (continue_polling, retry_merchant_request_with_operator_token, use_stored_operator_token, create_new_session, verification_failed, contact_support).

session = client.create_session()
print(session["verify_url"], session["poll_url"], session["poll_secret"])
print(session["next_steps"]["action"])  # "deliver_verify_url_and_poll"

status = client.poll_session(session["session_id"], session["poll_secret"])
if status["status"] == "verified":
    print(status["operator_token"])  # "opc_...", use for future requests

# Optional pre-association: attach the session to a known wallet or refresh KYC
# for an existing operator credential.
client.create_session(address="0x...")
client.create_session(operator_token="opc_...")  # KYC refresh

Wallet resolution

assess() responses include resolved_operator and linked_wallets, all same-operator sibling wallets (claimed via SIWE or captured via prior associate_wallet). The list may mix EVM addresses (0x... lowercased) and Solana addresses (base58, case-preserved) for cross-chain operators; merchants doing wallet-signer-match checks should accept a payment signed by any address in the list, regardless of chain. The address parameter on assess() and get_reputation() accepts either format; the network is auto-detected from the address shape.

Server-side signer-match + sanctions screening

Pass signer={"address", "network"} on assess() / aassess() to opt into server-side wallet-signer-match and OFAC SDN wallet-address screening on the same call. The response carries two new verdicts:

result = client.assess(
    "0xclaimed...",
    signer={"address": "0xsigner...", "network": "evm"},
    policy={"require_sanctions_clear": True},
)

# signer_match: wallet-binding verdict
#   kind: 'pass' | 'wallet_signer_mismatch' | 'wallet_auth_requires_wallet_signing'
#   plus claimed_operator / signer_operator / expected_signer / actual_signer /
#   linked_wallets / agent_instructions
match = result.get("signer_match")
if match and match.get("kind") == "wallet_signer_mismatch":
    # signer wallet resolves to a different operator than the claimed address
    ...

# signer_sanctions: OFAC SDN wallet-address verdict (discriminated union)
#   {"status": "clear"} | {"sanctioned": True, "ofac_label", "sdn_uid", "listed_at"}
#   | {"status": "unavailable"}
sanctions = result.get("signer_sanctions")
if sanctions and sanctions.get("sanctioned"):
    print("OFAC hit:", sanctions["ofac_label"], sanctions["sdn_uid"])

Wallet-OFAC SDN enforcement on the signer block is unconditional whenever a signer is supplied; no policy.require_sanctions_clear opt-in is required. The API flips decision to deny when signer_sanctions is sanctioned: True OR status: "unavailable", with decision_reasons including sanctions_flagged or sanctions_check_unavailable respectively (fail-closed; OFAC strict-liability). policy.require_sanctions_clear is the separate NAME-based screen on the resolved operator's KYC identity.

Pass signer["address"] = None for rails without a wallet signer (Stripe SPT, card-only). The API responds with signer_match["kind"] == "wallet_auth_requires_wallet_signing" and a parsed agent_instructions block telling the agent to switch to X-Operator-Token auth, spread the block directly into a 403 body.

Credential Management

cred = client.create_credential(label="my-agent", ttl_days=7)
print(cred["credential"])  # shown once

credentials = client.list_credentials()
client.revoke_credential(cred["id"])

Report an Agent's Wallet (Cross-Merchant Attribution)

After an agent authenticated via operator_token completes a payment, report the signer wallet so AgentScore can build a cross-merchant credential↔wallet profile. Fire-and-forget, first_seen is informational only. network is the key-derivation family: "evm" for any EVM chain (Base, Tempo, Ethereum, …) or "solana" for Solana.

client.associate_wallet(
    operator_token="opc_...",
    wallet_address=signer_from_payment,  # e.g. EIP-3009 `from` or Tempo MPP DID address
    network="evm",
    idempotency_key=payment_intent_id,  # optional, agent retries of the same payment no-op
)

Async

All methods have async variants prefixed with a:

async with AgentScore(api_key="as_live_...") as client:
    rep = await client.aget_reputation("0x1234...")
    result = await client.aassess("0x1234...", policy={"require_kyc": True})

    # Identity model methods
    session = await client.acreate_session()
    status = await client.apoll_session(session["session_id"], session["poll_secret"])
    cred = await client.acreate_credential(label="my-agent")
    await client.alist_credentials()
    await client.arevoke_credential(cred["id"])
    await client.aassociate_wallet(
        operator_token="opc_...",
        wallet_address="0x...",
        network="evm",
    )

Context Manager

with AgentScore(api_key="as_live_...") as client:
    rep = client.get_reputation("0x1234...")

Configuration

Parameter Default Description
api_key (required) API key from agentscore.com; raises ValueError if missing
base_url https://api.agentscore.com API base URL
timeout 10.0 Request timeout (seconds)
user_agent None Prepended to the default User-Agent as "{user_agent} (agentscore-py/{version})". Use to attribute API calls to your app.

AgentScoreError.status is a property aliasing .status_code so polyglot codebases can use the same attribute name regardless of which SDK raised the error.

Error Handling

from agentscore import AgentScore, AgentScoreError

try:
    rep = client.get_reputation("0xinvalid")
except AgentScoreError as e:
    print(e.code, e.status_code, str(e))

AgentScoreError.details carries the rest of the response body (verify_url, linked_wallets, claimed_operator, actual_signer, expected_signer, reasons, agent_memory) so callers can branch on granular denial codes without re-parsing.

Typed error classes

For status-code-specific recovery, the SDK raises typed subclasses of AgentScoreError. All inherit from AgentScoreError so existing except AgentScoreError still catches them.

Class Triggered by What it adds
PaymentRequiredError HTTP 402 The endpoint is not enabled for this account
TokenExpiredError HTTP 401 with error.code = "token_expired" Parsed body fields on the instance: verify_url, session_id, poll_secret, poll_url, next_steps, agent_memory, recover without re-parsing details
InvalidCredentialError HTTP 401 with error.code = "invalid_credential" Permanent, switch tokens or restart
QuotaExceededError HTTP 429 with error.code = "quota_exceeded" Account-level cap reached; don't retry
RateLimitedError HTTP 429 with error.code = "rate_limited" Per-second sliding-window cap; retry after Retry-After
TimeoutError httpx.TimeoutException (connect/read/write/pool timeout) Distinct from generic network errors. Note: subclasses AgentScoreError, not the builtin TimeoutError, import explicitly from agentscore.errors to disambiguate.

All non-timeout httpx.HTTPError (ConnectError, ProtocolError, NetworkError, etc.) are wrapped as AgentScoreError(code="network_error", status_code=0).

from agentscore import (
    AgentScore,
    AgentScoreError,
    TokenExpiredError,
    QuotaExceededError,
)
from agentscore.errors import TimeoutError as AgentScoreTimeoutError

try:
    client.assess("0xabc...", policy={"require_kyc": True})
except TokenExpiredError as e:
    print("Verify at:", e.verify_url, "poll with:", e.poll_secret)
except QuotaExceededError as e:
    print("Account quota reached, surface to user; don't retry.")
except AgentScoreTimeoutError:
    print("Network timeout, retry with backoff.")
except AgentScoreError as e:
    print(e.code, str(e))

Quota observability

assess() (and aassess()) responses include an optional quota field captured from X-Quota-Limit / X-Quota-Used / X-Quota-Reset response headers. Use it to monitor approach-to-cap proactively (warn at 80%, alert at 95%) before a 429:

result = client.assess("0xabc...", policy={"require_kyc": True})
quota = result.get("quota")
if quota and quota["limit"] and quota["used"]:
    pct = (quota["used"] / quota["limit"]) * 100
    if pct > 80:
        print(f"AgentScore quota at {pct:.1f}%, resets {quota['reset']}")

quota is absent when the API doesn't emit the headers (Enterprise / unlimited tiers). On a 429 response the SDK raises QuotaExceededError / RateLimitedError instead of returning a body, so quota is only readable on successful calls; drive proactive alerting off the success-path field.

Telemetry

telemetry_signer_match(payload) and atelemetry_signer_match(payload) are fire-and-forget POSTs to /v1/telemetry/signer-match so AgentScore can track aggregate signer-binding behavior across merchants. Used internally by agentscore-commerce's gate; available directly for custom integrations that perform their own wallet-signer-match checks.

Documentation

License

MIT

Metadata

Release files for agentscore-py 2.6.7

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for agentscore-py 2.6.7
File Size Uploaded
agentscore_py-2.6.7.tar.gz 67.6 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for agentscore-py 2.6.7
File Interpreter ABI Platform
agentscore_py-2.6.7-py3-none-any.whl Python 3 none any Details

Total release size: 89.9 kB

Release files / agentscore_py-2.6.7.tar.gz

Download URL agentscore_py-2.6.7.tar.gz
Size 67.6 kB
Tags Source
SHA-256 checksum
How to use checksums
e042118046fefa3ccdf1f9e0059b4adb489e4654adc8355b7a7105b422141d94
BLAKE2b-256 checksum
How to use checksums
50ede17e40adefb78114124dce1d9308a4a6c1bd8e374c1949940744e97385bd
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 31, 2026.

Transparency log

Release files / agentscore_py-2.6.7-py3-none-any.whl

Download URL agentscore_py-2.6.7-py3-none-any.whl
Size 22.4 kB
Tags Python 3
SHA-256 checksum
How to use checksums
4688f62d7d5d9848a6b1e42d3bee275a17ddf71342bef65b1bdf5d7177327a23
BLAKE2b-256 checksum
How to use checksums
bdf1b72dea059cd64e3133886f9aef32d335c7d37d611c36e4a49e1938c5c6b1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

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.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 31, 2026.

Transparency log

Release history Release notifications | RSS feed

2.7.1

2 release files

2.7.0

2 release files

2.6.11

2 release files

2.6.10

2 release files

2.6.9

2 release files

2.6.8

2 release files

This release

2.6.7 This release

2 release files

2.6.6

2 release files

2.6.5

2 release files

2.6.4

2 release files

2.6.3

2 release files

2.6.2

2 release files

2.6.1

2 release files

2.6.0

2 release files

2.5.3

2 release files

2.5.2

2 release files

2.5.1

2 release files

2.5.0

2 release files

2.4.3

2 release files

2.4.2

2 release files

2.4.1

2 release files

2.4.0

2 release files

2.3.4

2 release files

2.3.3

2 release files

2.3.2

2 release files

2.3.1

2 release files

2.3.0

2 release files

2.2.1

2 release files

2.1.2

2 release files

2.1.1

2 release files

2.1.0

2 release files

2.0.2

2 release files

2.0.1

2 release files

2.0.0

2 release files

1.8.0

2 release files

1.7.0

2 release files

1.6.0

2 release files

1.5.0

2 release files

1.4.2

2 release files

1.4.1

2 release files

1.4.0

2 release files

1.3.0

2 release files

1.2.0

2 release files

1.1.0

2 release files

1.0.4

2 release files

1.0.3

2 release files

1.0.2

2 release files

1.0.1

2 release files

1.0.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page