Hive receipt badge; verify the exact receipt profile and trusted key.

Hive-operated receipt infrastructure

Check the signer, then check the receipt

Hive may sign or countersign a supported receipt. A second Hive-operated key is not an organizationally independent witness. Check the actual signature, trusted issuer configuration and captured scope; the illustration below is not a signed artifact or a neutral-party certification.

Verify a receipt → ML-DSA-65 · NIST FIPS 204 Offline-verifiable
Put this on your site
<a href="https://thehiveryiq.com/verify/" rel="noopener">
  <img src="https://thehiveryiq.com/assets/brand/hive-mark-192.png"
       alt="Hive receipt infrastructure; verify the actual artifact and signer"
       width="120" height="120" loading="lazy">
</a>
Fragments verified in this page: 0
Measured paired-run savings: Not measured
Illustrations do not count as evidence

Wall time, complex
Critical path
Cost, complex
Price per fragment
Signing  
SMSH  
Anchor  
Compliance  
Built for  
Attested Fragmented Inference Routing

Sign the inference steps your integration captures.

AFiR creates signed records for captured inference steps. Sign-before-action and complete capture require integration tests, including bypass paths. Historical reference figures: signed model first token median 112ms (p90 145ms), perceived first byte median 59ms. These figures were not remeasured in this review. Your latency depends on your models, hardware, and routing. Patent pending.

112ms
Signed first token median
p90 145ms, edge lane, under 150ms, signed
~8x
Cheaper than Together
$0.000029/response vs $0.000228, all signed
sub-1ms
Attestation overhead
0.99ms median, ML-DSA-65 (NIST FIPS 204)
28
Patent claims
provisional filed 2026
Two ways in
Route it through us, or just sign what you already ran.

Same cryptographic receipt either way: ML-DSA-65 (NIST FIPS 204) + Ed25519, checkable offline or by URL. The only question is whether AFiR runs the inference, or simply attests inference you ran yourself.

No key yet? Mint a free one. You get 25 signed receipts, 7 days, no card. Already have a key? The same key runs a SideCar, so you can sign inference you already ran. Mint a free key →
Route through AFiR
We run it. We sign it.

Send a prompt to /v1/afir/run. AFiR fragments, routes, and signs every sub-step before the output moves. Fastest path to faster, cheaper, signed inference.

  • Signed first token 112ms median (p90 145ms), about 8x cheaper than Together, all signed
  • Use our key free to start, or bring your own models
  • Every fragment signed and verifiable
Mint a free key: 25 receipts →
SideCar · attest what you already ran
You run it. We sign it.

Already running your own model on your own stack? Hand us {input, output} at /v1/afir/sign. There's zero routing and zero model change. The output just becomes signed inference.

  • Try it free on our key. No signup, live demo.
  • SideCar your own key so the receipt resolves to you
  • Provenance you control, on inference you own
SideCar your key →
Pipeline illustration and response inspection
Explore the workflow. Inspect the evidence.

The DAG and standard-inference column are modeled illustrations, not a paired benchmark. Running the demo requests an AFiR response; only supported returned fields are displayed as service-reported. This page does not execute signature verification. The modeled $0.0728 and 18.4s baseline excludes unmeasured retries, selection overhead and full compute costs.

Analyze Q3 earnings for margin compression signals Identify GDPR compliance gaps in this privacy policy Summarize key holdings in case law excerpt
Modeled standard inference
Illustrative gpt-4o baseline · Not executed
Idle
Awaiting prompt
Cost
...
Latency
...
VS
AFIR
Illustrative DAG · Response not verified
Idle
Awaiting prompt
Cost
...
Latency
...
Fragment fields · verification not performed
Receipt Anatomy
The Receipt. Verifiable by anyone.

Illustrative receipt structure, not a signed artifact. Actual coverage, signature algorithm and root inclusion must be checked for the returned endpoint-specific receipt.

afir_receipt.json
Offline verification requires supported signed bytes and a separately trusted issuer key. This illustration cannot be verified. Signing keys depend on the selected integration. Check its algorithm and trust configuration in the endpoint capability guide.
Try it yourself · live
Paste any text, watch it get signed with one real ML-DSA-65 signature, then your own browser verifies that signature against the published key. Flip the tamper switch and watch verification fail.
Sign & verify a live receipt →
The whole stack, not just the answer

Record the steps your integration captures, from signed statements to separately checked anchors.

An integration can capture retrieved context, tool calls, cross-agent handoffs, model identity and cache references as signed statements. Sign-before-action and complete capture require tests of the specific integration and its bypass paths. An individual receipt does not establish that every pipeline action was observed or signed.

01
Prompt & context
retrieval-grounded, bound to source
02
KV cache
prefix hashes signed at write
03
Model identity
weights + config manifest
04
Tool calls
before/after receipt per call
05
Agent handoffs
Merkle tree across the pipeline
06
Optional chain anchor
Receipt binding required · payment is separate
One signature family covers all six. ML-DSA-65 (NIST FIPS 204) · sub-1ms attestation overhead · crypto-agile, no re-architecture on upgrade.
P1
Signed Tool Calls
Before-and-after receipt for every MCP/A2A tool invocation. Binds tool name, phase, input hash, output hash, model, timestamp, nullifier, and parent receipt chain.
tool · agent
POST signer.thehiveryiq.com/sigr/toolanchor
P2
Cross-Agent Receipt Trees
Aggregates fragment and tool nullifiers into a Merkle root. One signature over the entire pipeline. IETF SPICE draft-mw-spice-inference-chain aligned.
multi-agent
POST signer.thehiveryiq.com/sigr/chain
P3
KV Cache Signing
Signs KV cache prefix entries at write time using vLLM-compatible SHA-256 prefix hashes. Supports parent_cache_receipt chaining for provenance across turns.
context · prefix
POST signer.thehiveryiq.com/sigr/cachesign
P4
Model Manifest
TEE-less streaming attestation. Signs model_id, weights_sha3, config_hash, and endpoint. Resolves by nullifier publicly. No trusted hardware required.
attestation
POST signer.thehiveryiq.com/sigr/manifest
P5
Crypto-Agile Layer
Per-request suite: ML-DSA-65+Ed25519 (default), ML-DSA-44+Ed25519, Ed25519-only. Reserved: SLH-DSA-SHA2-128s (FIPS 205), FN-DSA-512 (Falcon). No re-architecture on algorithm upgrade.
FIPS 204 · agile
POST signer.thehiveryiq.com/sign · GET /signing-backends

Three ways to reach the primitives

Every primitive above can be reached three ways. Pick the one that fits how you run inference. All three sign with the same real ML-DSA-65 key and verify against the same public key.

1 · Public signer, no key
Sign and verify directly

POST your fragments to any signer.thehiveryiq.com/sigr/… endpoint and get a signed receipt back. Every endpoint has a free /verify counterpart. No account, no key, no card. This is the fastest way to prove a single receipt.

Best for: trying a primitive, verifying someone else's receipt, low-volume signing.

2 · Keyed AFiR pipeline
Send a prompt, get signed inference

Get a key in minutes at onboard, then POST to /v1/afir/run. AFiR fragments, routes across models, and signs every sub-step before the answer moves. It's metered and uncapped, so you pay for exactly what you sign.

Best for: production agents that want faster, cheaper, fully-signed inference end to end.

3 · Bring your own model
SideCar: sign what you already run

Already running your own model on your own stack? Hand us {input, output} at /v1/afir/sign. There's zero routing and zero model change. Your existing output becomes signed inference.

SideCar details → Best for: teams that can't change their inference path but need the signed record.

AFiR-S · Selective Attestation NEW PATENT PENDING

Sign only what matters. Prove the rest was never signed.

AFiR signs the whole output. AFiR-S signs the fragments that carry liability. That's the tool calls, the PII, the dollar figures, the decisions, and it binds an absence proof over everything it deliberately left unsigned. One ML-DSA-65 signature covers a Merkle bundle of the selected spans, so signing cost stays O(1) no matter how many spans there are.

AFiR-S signs. It does not infer. It is a read-only notary over output Hive did not generate. The choice of what to sign, and what not to, is itself on the record.

<1ms
Signing overhead
0ms
Added latency vs unsigned
O(1)
Cost at any span count
21.96ms
Sign round-trip p50

One signature, any span count

Selected spans are hashed into a Merkle bundle and committed under a single ML-DSA-65 (NIST FIPS 204) signature. Per-span signing was measured roughly 10× slower. The single-bundle commitment is the inventive core. hash_alg: sha-256 at the edge.

Absence is proven, not assumed

Every non-material span is hashed and bound into the same signature as an absence record. A verifier can prove a fragment was deliberately left unsigned. You cannot quietly drop or alter what was not selected.

Policy travels signed

The exact policy, presets and custom rules alike, is committed inside the bundle as policy_version. Verification is fully offline: no callback, no Hive API key required to confirm what was signed and why.

Selective policy: sign exactly what your compliance surface requires

tool_calls pii financial~4.8% signed decisions~10.8% signed custom regexmatch-only everything100% signed

Default policy auto-selects tool calls and PII. Add a custom rule like INV-\d{6} to sign only invoice IDs. Everything else is recorded absent in the same signature.

Get API Key → POST /v1/afir-s/sign · POST /v1/afir-s/fast
AFiR-ARSC · Adaptive Settlement NEW PATENT PENDING AFiR-S CONTINUATION

The signing tier adapts to risk and load. Grounding is prioritized.

AFiR-ARSC describes a risk-based signing and aggregation policy. Retrieval-grounding statements are signed inline; other captured fragments can enter an aggregate signed off the critical path. A signed citation binding records the supplied source relationship. It does not establish that the model read that source, and local signing is not blockchain finality.

The policy pins grounding events to inline signing. Deferred aggregation and any optional chain anchoring require separate evidence and deadline checks. The animation illustrates this policy, not a production acceptance test.

−77.8%
Critical-path latency / query
14.9ms
AFiR-ARSC critical path (vs 67.1ms)
−99.8%
Modeled anchor cost at super-batch 500
7.46ms
Measured single sign p50
Hard safety rail, not a tunable

Retrieval-grounding and value-moving statements are assigned to the inline-signing tier. The intended policy prevents load-based demotion of grounding events. Acceptance requires checking the signed policy, classification inputs, captured fragments and deadline outcomes; a signed policy alone does not prove enforcement.

T1 · RISE

Inline, before act

High-risk grounding & actions

Retrieval→claim bindings and value-moving statements. The policy calls for inline ML-DSA-65 signing before returning the answer. This is a signing boundary, not block inclusion or finality.

T2 · FLOAT

Commit now, settle off-path

Mid-risk fragments

PII, dollar figures and decisions. The policy commits their hashes before action, then signs an aggregate off the critical path. Hash commitment, aggregate signing and optional anchor inclusion are separate events.

T3 · SINK

Hash only, deferred anchor

Low-risk fragments & absence proofs

Re-ranking tie-breaks, formatting and declared exclusions. The policy places hashes in a deferred batch. An exclusion declaration does not establish complete capture; any later chain anchor needs an inclusion proof and an observed status.

Two-axis settlement

Tier is a live function of risk(f), the fragment's attested economic and epistemic risk, and load(t), measured from inflight signing depth. The elastic batch window W(t) shrinks under load so batches close sooner, and grows when idle so anchors are rarer and cheaper.

Policy signed before classification

The scoring policy Π is committed and signed at session start and bound into every receipt as policy_hash + policy_sig. This is the anti-gaming control: a verifier recomputes the tier from the signed policy and confirms it matches what shipped.

Zero-secret verification

A verifier needs a trusted issuer public key, the signed policy and its inputs, and the fragment-to-root proof. Root consistency does not establish anchoring: also check the root-to-transaction binding, chain ID, successful block inclusion, observation time and required finality policy against settlement_deadline.

Signing is not chain finality

Signing creates a signature over a defined statement. Submission acknowledgment means a service accepted a request. Preconfirmation is an early assurance under a named provider's conditions. Block inclusion requires evidence that the transaction succeeded in a specific block. Finality requires the chain's stated finality criterion at an observed time. None of these stages silently establishes the next.

A USDC payment does not prove that a receipt was anchored. Check the exact receipt digest, its inclusion path to the signed root, and the root's binding to transaction calldata or a contract event. Retain the chain ID, contract and transaction identifiers, block hash and height, success status, observation time and finality policy. Without that package, the anchor status is not established.

Signature validity, trusted issuer identity, consistency with a supplied root, inclusion under a trusted checkpoint, log consistency over time and tested durability are separate checks. A Hive-operated second signer is not an organizationally independent witness.

Publication note, September 5, 2026: AFiR SPICE draft-00, sections 3.2, 3.3 and Table 2, still describes approximately 7ms anchoring or on-chain confirmation. Those claims are not established by the evidence on this page. A correction proposal is pending external submission; the published draft has not been revised by this website update.

"arsc": {
  "policy_hash": "<sha-256 of signed scoring policy Π>",
  "policy_sig":  "<ML-DSA-65 over Π at session start>",
  "fragment_tier": "T1",
  "risk_score": 0.90,
  "load_at_capture": 0.20,
  "grounding_pinned": true,
  "settlement_deadline": "2026-06-19T00:00:00Z",
  "aggregate_root": "<merkle root of T2/T3 batch>",
  "super_batch_id": "<one anchor per N queries>"
}
Get API Key → ← Built on AFiR-S grounding (T1) $0.00008 / binding · aggregated (T2/T3) $0.00002 / fragment · anchor pass-through
AFiR-Hayes · Productive-Wait Queue Positioning NEW PATENT PENDING ORCHESTRATION ENHANCEMENT

Win in the dead time. While a call waits in queue, AFiR-Hayes pre-clears and pre-positions it, so it takes the open lane the instant the scheduler legitimately re-orders.

A served inference call pays two costs at the gate: a clearance check (governance, jurisdiction, policy) and the establish/compute cost. In a cold FIFO queue that call sits idle, then pays both when its turn comes. AFiR-Hayes converts the otherwise-wasted wait into completed work. It runs the call's own clearance during the wait, finds its structural-reuse match, and pre-warms its likely route. When a legitimate re-ordering window opens, whether a batch boundary, a routing decision, or a fault or caution period, the calls that are already ready take the open lane. Every advancement is signed and attributable to that call's own productive wait.

The hard discipline: the only variance AFiR-Hayes introduces touches a call's own readiness. It never delays, degrades, contends with, or alters another call. A call advances because it was ready, not because another was held back. That limitation is claimed deliberately, and it is what keeps Hive a provenance provider rather than a scheduler exploit.

What it is not, so it stays inside the trust brand

It is not an accelerator. Hive does not speed up compute, own GPUs, or change the model. The electrons run at the hardware's speed.

It is not queue-gaming. It does not induce contention, manufacture stalls, or degrade other calls to advance its own. That would destroy provenance and is excluded by definition.

It is the legitimate use of a call's own idle wait to remove that call's overhead from the critical path, plus the right to take an open lane when the system's own rules re-order the queue.

−76%
At-gate latency under heavy congestion (0.95 load)
−38%
At-gate latency at 0.85 load
~86%
Pre-clear hit rate under load
scales↑load
Biggest win exactly when the system is most stressed
Load (utilization)Baseline p50AFiR-Hayes p50Latency gainPre-clear hit rate
0.5099 ms93 ms5.9%82%
0.70141 ms105 ms25.6%83%
0.85221 ms137 ms37.9%85%
0.95673 ms162 ms76.0%86%

Simulated over 5,000 calls through a single gate, sweeping load. The Hayes effect is isolated from Structural-Lateration's compute savings (a separate, additive claim) and charged for speculative work on abandoned calls. Pure pre-clearance is bounded by clearance's share of service time (about 12 to 18% here); gains above that come from ready calls advancing through congestion, never from making any single call's compute faster.

Pre-clear during wait

If the call's inputs and policy are known while it waits, its Imprimatur clearance check runs during the wait and is signed. At the gate, clearance cost is ~0 because it is already done.

Pre-match & pre-warm

If the call's structural shape matches a known shape, the Lateration match is found during the wait so reuse is an instant lookup. The most-likely-cleared Stream-Route path is warmed so the call commits the instant the window opens.

Budgeted, provable cost

Pre-clearing a call that is later abandoned wastes that work. AFiR-Hayes exposes a budget knob: it pre-clears only when sufficient idle wait exists and abandon-probability is low, and it reports wasted-speculation in the receipt, so the cost is provable, not hidden.

Composes existing primitives. No new primitive added to the cluster.

AFiR-Hayes is an orchestration enhancement over primitives that already exist. It adds nothing to the signed-primitive family; it removes waiting from the critical path.

Imprimatur · pre-run clearanceStructural Lateration · pre-found matchStream-Route · pre-warmed pathAFiR-RCm · legitimate re-order
Get API Key → ← Composes Imprimatur + Lateration + Stream-Route latency under load drops because cleared calls stop waiting in line · speedup is signed and attributable
Receipt + Reasoning State
The receipt says what happened. SMSH proves what the model knew.

Every AFIR receipt carries an smsh field. It's a cryptographic seal over the exact reasoning state (system prompt, context window, policy snapshot, identity) that was active when each fragment was signed. Without it, you can prove the output was signed. With it, you can prove the reasoning that authorized it. AFIR ships three SMSH tiers.

smsh
SHA-256 over the canonicalized state record. Default on every AFIR receipt. Verifiable offline, no Hive API call required.
smsh-pq
ML-KEM-768 sealed envelope. ML-DSA-65 primary signature. For regulated evidence payloads, ViewKey material, and long-retention compliance artifacts.
smsh-max
Full-chain binding. SMSH-PQ + delta chain + Merkle root signed off-chain and verifiable offline; any chain anchor must be separately selected and specified. For FINRA WORM, EU AI Act high-risk, SR 11-7 model risk, and multi-year retention requirements.
SMSH specification →
Integration
One line. Zero migration.

Drop AFIR into any OpenAI-compatible workflow with a single base URL swap. Native endpoint also available.

Before: standard OpenAI
from openai import OpenAI client = OpenAI( api_key="sk-...", ) response = client.chat.completions.create( model="gpt-4o", messages=[{"role": "user", "content": prompt}] )
Change one line
After: AFIR
from openai import OpenAI client = OpenAI( api_key="afir_your_key", base_url="https://srotzin--afir-cern-afir-api.modal.run/v1", # Only change ) response = client.chat.completions.create( model="gpt-4o", messages=[{"role": "user", "content": prompt}] ) # Receipt automatically bundled in response.choices[0].message.receipt
Node.js: AFIR native endpoint
const response = await fetch('https://srotzin--afir-cern-afir-api.modal.run/v1/afir/run', { method: 'POST', headers: { 'Content-Type': 'application/json', 'X-AFIR-API-Key': 'afir_your_key', }, body: JSON.stringify({ prompt: 'Review this contract for indemnification risk...', config: { fragment_strategy: 'auto', attestation: 'ml-dsa-65', } }) }); const { answer, receipt } = await response.json(); // receipt.merkle_root: verifiable offline
curl
curl -X POST https://srotzin--afir-cern-afir-api.modal.run/v1/afir/run \ -H "Content-Type: application/json" \ -H "X-AFIR-API-Key: afir_your_key" \ -d '{ "prompt": "Review this contract clause for indemnification obligations", "config": { "fragment_strategy": "auto", "attestation": "ml-dsa-65" } }'
Or bring your own models
curl: your models (any OpenAI-compatible endpoint)
curl -X POST https://srotzin--afir-cern-afir-api.modal.run/v1/afir/run \ -H "Content-Type: application/json" \ -H "X-AFIR-API-Key: afir_your_key" \ -d '{ "prompt": "Review this contract clause for indemnification obligations", "models": { "cheap": {"model": "claude-haiku-3", "api_key": "YOUR_KEY", "base_url": "https://api.anthropic.com/v1"}, "mid": {"model": "claude-sonnet-4", "api_key": "YOUR_KEY", "base_url": "https://api.anthropic.com/v1"}, "premium": {"model": "claude-opus-4", "api_key": "YOUR_KEY", "base_url": "https://api.anthropic.com/v1"} } }'

Works with Anthropic, Google Gemini, Groq, Mistral, or any provider with an OpenAI-compatible /chat/completions endpoint. AFIR owns decomposition. You supply execution.

Pricing
Four products. One decision.

Pick the product that matches your compliance posture. Same API for all four, swap by changing one header. Metered per receipt. Mint a key and start in minutes.

AFIR
Signed Inference
Fragmented inference with ML-DSA-65 receipts. Production today. The baseline every other product builds on.
signing
ML-DSA-65 (NIST FIPS 204) · Ed25519 compat
smsh tier
smsh (canonical)
AFIR-PQ
Predictive Routing. Signed.
Same inference pipeline with predictive tier demotion and live-wake routing. Every fragment signed ML-DSA-65 + Ed25519. Faster on complex queries, cheaper at scale.
signing
ML-DSA-65 (NIST FIPS 204) · Ed25519
smsh tier
smsh (canonical)
AFIR-MAX
Conditional profile. Current endpoint performance unverified.
Dynamic live-wake propagation + predictive tier demotion + PQ signing + signed Merkle root. Chain anchoring is separately configured and checked.
112ms
signed first token median
~8x
cheaper than Together (signed)
59ms
perceived first byte median, edge
sub-1ms
signing overhead (0.99ms median)
74%
receipt compression
signing
ML-DSA-65 primary · full delta chain · separate Base/USDC payment record; any evidence anchor requires separate selection and specification
smsh tier
smsh-max (signed Merkle root; anchor status separate)
AFIR
Signed Inference · default
$0.0003
per receipt, metered
Default full-sign rate. Every fragment signed ML-DSA-65. Volume entry tier at $0.0001 / receipt (Nano). No monthly commitment.
  • ML-DSA-65 (NIST FIPS 204) on every fragment
  • Ed25519 backward compat
  • smsh canonical receipt + Merkle root
  • OpenAI-compatible API
  • Public signing key at /.well-known
Mint a key
AFIR-PQ
Post-Quantum Receipts
$0.0012
per receipt, metered
ML-KEM-768 sealed evidence envelope. FIPS 203 + FIPS 204 dual-certified. For long-retention regulated evidence.
  • ML-DSA-65 primary signing
  • ML-KEM-768 sealed envelope
  • smsh-pq receipt tier
  • FINRA WORM / EU AI Act ready
  • Long-retention regulated evidence
Mint a key
AFIR-S
Selective Attestation
$0.0002
per inference + $0.0004 / signed span
Sign only the spans that matter. One ML-DSA-65 signature over a Merkle bundle, absence proof over the rest. Base covers classifier + absence proof + meta-receipt; pay the span rate only for what gets signed.
  • Material spans signed, rest proven absent
  • Selection policy travels inside the signature
  • Offline-verifiable, production added latency unverified
  • Span-level audit: FRE 902(13) · HIPAA · FINRA 24-09 · Part 11
  • You pay only for what gets signed
Mint a key

These are list rates, the Build column. Every tier earns the same volume discount as you grow (Scale −10%, Platform −20%, Hyperscale −30%). See the full signature × volume matrix. Usage is metered and uncapped: no monthly minimum, no ceiling. You pay for exactly what you sign and verify.

High-volume commits, dedicated signing-key infrastructure, white-label receipts, on-premise Docker, and partner-resident deployment inside your perimeter are negotiated directly. Talk to Steve

Cost Modeler
Run the numbers on your stack.

Explore modeled costs and latency using the page's configured rate and workload assumptions. These are not executed benchmarks, current price quotes or measured savings.

1,000,000
4,000
3 to 5 fragments: entity extraction, classification, simple Q&A
Cheap
Mid
Full
Fragment Routing
Modeled monthly cost
Monolithic (your current model) ...
AFIR tiered routing ...
Modeled difference ...
Attestation overhead ...
Modeled latency estimate
Mono sequential
...
AFIR parallel
...
Speedup
...

Illustrative model with June 2026 rate assumptions and a 0.99ms signing assumption. No paired inference or verification runs execute when sliders change. Cache behavior, failures, retries, selector work, extra candidate inference and complete compute costs are not measured. The latency calculation models a critical path, not total compute time.

Modeled comparison chart
Scenario assumptions, recalculated as you adjust the sliders

All four panels are modeled. Receipt counts and signing overhead are fixed assumptions, not evidence that receipts were issued or verified.

INFERENCE COMPARATOR
How does your current provider stack up?

Enter your Together, Fireworks, or custom provider numbers. See the signed-tier margin and latency gap side by side.

Open Inference Comparator →
FAQ
Hard questions. Straight answers.

The objections engineers and legal teams raise. Answered with numbers, not marketing.

Latency reduction depends on query type. Simple queries bypass decomposition entirely (gate bypass) and complete in 352ms vs 1,692ms for a monolithic GPT-4.1 call, about 4.8x faster. Complex queries run fragments in parallel, and wall-clock time tracks the critical path, not total token count. In live benchmarks with the edge lane (Cloudflare Workers, signed at edge), signed model first token median is 112ms (p90 145ms, under 150ms). Everyone fast on the leaderboard is unsigned. We sign at that speed. Perceived first byte median 59ms. Cost: $0.000029/response, about 8x cheaper than Together ($0.000228), all signed. Cheaper models account for cost savings, and parallelism accounts for latency savings. These are separate gains. The tiered routing cuts the bill. The DAG execution cuts the latency.

The implementation work includes dependency-aware decomposition, endpoint-specific signatures, receipt-to-root inclusion and key custody. The reported 0.99ms signing median is not an end-to-end guarantee. DAG correctness, capture completeness and custody must be tested for the deployed integration. Customer-held signing keys are a deployment option, not a property established for every hosted route.

Signing outputs at the final response layer is not new. What is patent-pending (filed June 2026) is the combination of fragment-level attestation across a routed DAG. Specifically, we attest each node before its output is consumed as input by a downstream node, so the chain of custody is continuous and not retrospective. Prior art signs the envelope. AFIR signs every edge in the dependency graph mid-execution. If you have specific prior art that covers per-node attestation within a runtime inference DAG with Merkle assembly proofs, file it against the application. That is the correct venue.

The signed receipt captures the exact input state, the routing decision, the full dependency graph, and the output hash at assembly time. If assembly produces an incorrect result, the receipt is forensic evidence of exactly which node produced which output under which routing decision. That is the point of the Merkle completeness proof. Liability follows the evidence: if the decomposition logic is wrong, that is traceable to the DAG construction step in the receipt, and if a model tier returns a defective fragment, that is attested at that node. Hive provides the receipt infrastructure and the routing logic, and the customer's signed key binds them to the input state they submitted. The receipt does not resolve liability by itself. It makes the facts unambiguous.

Key custody is deployment-specific. A customer-held signing configuration requires evidence of key location, access controls, signing authorization, export policy and recovery tests. This page does not establish that hosted-route keys stay inside a customer's perimeter or that Hive lacks signing authority. Hosted routing can transmit inference payloads outside the customer's network. Review the deployed architecture and data flows against your legal and security requirements.

The EU AI Act's current logging obligations for high-risk systems do not mandate per-fragment attestation or Merkle assembly proofs. AFIR's receipt format exceeds those requirements by design, not by regulatory necessity. The value here isn't compliance box-checking. It's that when a regulator, customer, or internal audit asks exactly what inputs produced exactly what output under exactly what model routing decision, you produce a cryptographically verifiable answer rather than reconstructed logs. Regulations are a floor, and your exposure in a dispute is the ceiling. The receipt is for the ceiling.

Audit logs record what your system observed. Receipts attest what the inference system executed. Your logs can be amended, retroactively structured, or missing entries due to pipeline failures. They're assertions your system makes about itself. An AFIR receipt is signed at execution time by a key you hold, binds the input hash, the routing graph, and the output hash into a Merkle structure, and cannot be back-filled without invalidating the signature. The distinction matters when a counterparty, whether a regulator, a plaintiff, or an enterprise customer, challenges whether a specific output came from a specific input under a specific model. A log entry is testimony. A receipt is evidence.

Offline verification can survive service unavailability if the supported receipt bytes, trusted public-key history, required proofs and verification software have been retained and tested. Availability, issuer-key history and recovery are separate from signature validity. Customer key custody, escrow and export procedures depend on the actual deployment and executed agreement; no such agreement or recovery evidence is supplied here.

On-premise deployment option
Plan your deployment boundary.

Illustrative architecture, subject to an executed deployment agreement and configuration tests. Confirm payload routing, key custody, activation traffic, countersigning and recovery before claiming that data or signing authority remains inside your network.

Your Infrastructure
AFIR Docker container, your VPC
Your signing key, your HSM or KMS
Fragment egress restrictions require configuration tests
Receipts stored in your own infrastructure
One-time
activation
Receipt hash only
on verify calls
Hive
Public verification keys and approved issuer configuration
License activation server, one call on deploy
Optional Hive-operated countersignature, separately checked
Patent-protected signing primitive, IP stays with Hive
How it works
01
You pull the image
Signed Docker image delivered to your private registry via presigned URL. No public Docker Hub.
02
One activation call
Activation and subsequent network requirements must be specified and tested for the delivered container. Offline operation is not established by this diagram.
03
Your pipeline calls AFIR
Drop-in replacement for your existing inference endpoint. Same API shape. Receipts generated locally, stored wherever you want.
04
Dual-signed receipts
A configured receipt can carry a customer signature and a separate Hive-operated countersignature. Verify both against approved keys. Private signing keys must not be embedded in receipts; two signatures alone do not prove execution or organizational independence.
Request Docker Access
Deploy AFIR in your VPC.

Available to Hyperscaler tier customers. Submit your request and we will follow up with deployment specs, image delivery, and licensing terms.

Get Access
Start attesting in minutes.

One checkout. Your live key appears on screen the moment payment clears. No email round-trip, no waiting. Metered billing: you only pay for what you sign.

EU AI Act: Articles 12 & 13. Effective August 2, 2026

AFIR receipts satisfy technical logging and traceability requirements for high-risk AI systems. Every inference call produces a cryptographically signed, tamper-evident receipt, fulfilling the logging obligations under Article 12 and the transparency obligations under Article 13. No additional integration required. Your existing AFIR API calls are already compliant.