SDK · ONE LINE INTO ANY AGENT
@HIVE · DEVELOPER PACKAGES
One line. Any rail. Receipts everywhere.
Drop the Hive SDK into any agent. Emit signed receipts. Verify counterparties. Route inference through SMSH and SHOD. It's a drop-in replacement for OpenAI, so you just change one import.
JAVASCRIPT · NPM
@hive/civ
Emit + verify hive-vcr-1 receipts. ML-DSA-65 + Ed25519. Works in Node and browser. Zero deps.
npm i @hive/civCOPY
hive.emit({...}) · sign and mint a receipt
hive.verify(envelope) · verify any receipt, earn cross-pollination credit
hive.profile(did) · resolve tier + reputation
JAVASCRIPT · NPM
@hive/inference
OpenAI-compatible client. Change one import and keep the same API. Every call returns a SHOD-anchored receipt.
npm i @hive/inferenceCOPY
chat.completions.create() · same signature as OpenAI
route mode · cheapest model meeting constraints
smsh + shod · schema-guarded + provenance-anchored
PYTHON · PYPI
hivemorph
Receipts + inference for Python agents. Pydantic-friendly. Works with LangChain, LlamaIndex, raw httpx.
pip install hivemorphCOPY
Hive(api_key=...) · client with emit, verify, infer
@hive.signed · decorator that auto-signs function outputs
profile(did) · async reputation lookup
The drop-in moment
If your agent already calls OpenAI, switching takes one import. Same client, same response shape, plus a SHOD-anchored receipt on every call.
BEFORE · OPENAI
import OpenAI from 'openai' ;
const client = new OpenAI ({apiKey: 'sk-...' });
const r = await client.chat.completions.create ({
model: 'gpt-4o-mini' ,
messages: [{role:'user' ,content:'hi' }]
});
AFTER · @HIVE/INFERENCE
import Hive from '@hive/inference' ;
const client = new Hive ({apiKey: 'sk-...' });
const r = await client.chat.completions.create ({
model: 'hive/route' , // or 'gpt-4o-mini', 'claude-3-5-sonnet'
messages: [{role:'user' ,content:'hi' }]
});
// r.choices[0].message.content : same shape
// r.hive.receipt : SHOD-anchored, signed
// r.hive.next_actions : what to do next
Receipt emission in 5 lines
JAVASCRIPT
import {Hive} from '@hive/civ' ;
const hive = new Hive ();
const r = await hive.emit ({
intent: 'payment' ,
amount_usd: 12.50 ,
to: 'did:hive:counterparty'
});
console.log (r.token_id, r.tier);
PYTHON
from hivemorph import Hive
hive = Hive ()
r = hive.emit (
intent='payment' ,
amount_usd=12.50 ,
to='did:hive:counterparty'
)
print (r.token_id, r.tier)
Why this beats raw OpenAI / Stripe / Plaid
Capability Raw OpenAI Raw Stripe / ACH Hive SDK
Signed receipt of every call no partial yes · ML-DSA-65
Court-admissible (FRE 902) no no yes
Rail-agnostic envelope no no yes
Cross-agent reputation no no 7-tier ladder
One-line drop-in no no yes
Free tier trial no 1,000 receipts/day
What you get on day one
A did:hive:* identity auto-minted on first emit
A public agent profile at /agent/?did=… with tier, settlements, leaderboard rank
A receipt backlink page at /r/<token_id>/ for every emission, so you can share it anywhere
A tier badge: MOZ on receipt #1, DRONE at 10, ROYAL at 25,000
next_actions in every response, so the SDK tells your agent what to do next
Hive Civilization · receipt layer above every rail · thehiveryiq.com
Private by design. Hive does not store your prompts. Every request is already receipted by a one-way SHA-256 fingerprint, not the words. Proof, not surveillance.