Skip to content
TAONIC

04 / 14VERIFIEDDescribes what runs today

Providers and subnet evidence

TAONIC runs no Bittensor infrastructure of its own. It is a customer of each provider below, through that provider’s public API. This page documents every route TAONIC has an adapter for; the live state of each is on Routes.

Every route

Every route TAONIC has an adapter for
RouteSubnetCredentialVerified
Text generation · ChutesSN64CHUTES_API_KEYNever
Text generation · engySN53ENGY_API_KEYNever
Text generation · OpenRouter to ChutesSN64OPENROUTER_API_KEY2026-09-24
Web search · DesearchSN22DESEARCH_API_KEYNever
Event forecasts · NuminousSN6None2026-09-24
Price forecasts · SynthSN50SYNTH_API_KEYNever
Image generation · ChutesSN64CHUTES_API_KEYNever
Text to speech · ChutesSN64CHUTES_API_KEYNever
Speech to text · ChutesSN64CHUTES_API_KEYNever
Embeddings · ChutesSN64CHUTES_API_KEYNever

Chutes, subnet 64

Chutes runs model inference on Bittensor subnet 64: miners serve the models and validators decide which miners serve. TAONIC reaches it two ways: through OpenRouter, pinned to Chutes with fallbacks off, and directly.

Text generation · OpenRouter to Chutes

VERIFIED
Route id
openrouter-chutes
Capability
Answer, Check a claim
Subnet
Bittensor subnet 64 (Chutes)
Connection
Through OpenRouter (intermediary): POST https://openrouter.ai/api/v1/chat/completions · model google/gemma-4-31b-it
Configuration
Server secret OPENROUTER_API_KEY. An OpenRouter API key from openrouter.ai/keys. The account is free to create; paid credit is needed for the Chutes endpoint, at Chutes’ own per-token rate.
Cost
Chutes’ own rate through OpenRouter: $0.12 per million input tokens and $0.37 per million output, about $0.0004 an answer, plus OpenRouter’s credit fee. Paid by TAONIC; checked 2026-09-16.
Provenance
OpenRouter names the provider that served each answer, and TAONIC refuses the answer unless that provider is Chutes. Neither OpenRouter nor Chutes publishes which miner hotkey served a request, so no answer is attributed to a particular miner.
Privacy
The question is sent to OpenRouter, then to Chutes, under their own terms.
Deadline
95 s
Evidence
Production request 30e55018-7fde-4f57-a7c6-b05056d7fe8d, 2026-09-24 05:16:47 UTC. OpenRouter reported Chutes as the serving provider; 298 input and 190 output tokens. Subnet 64 on taostats (opens in a new tab)
Documentation
OpenRouter to Chutes documentation (opens in a new tab)

Text generation · Chutes

ADAPTER READY
Route id
chutes-chat
Capability
Answer, Check a claim
Subnet
Bittensor subnet 64 (Chutes)
Connection
Direct: POST https://llm.chutes.ai/v1/chat/completions · model google/gemma-4-31B-turbo-TEE
Configuration
Server secret CHUTES_API_KEY. A Chutes API key (cpk_…) from chutes.ai/app/settings/api-keys, on a pay-as-you-go balance (Chutes has no free tier).
Cost
Gemma 4 31B (TEE) at $0.12 per million input tokens and $0.37 per million output tokens: about $0.0004 an answer. Paid by TAONIC; checked 2026-09-15.
Provenance
Chutes returns its own completion id, the model that answered and, when it sends them, invocation and instance ids. It does not publish which miner hotkey served a request, so no answer here is attributed to a particular miner.
Privacy
The question is sent to Chutes under its own terms.
Deadline
45 s
Evidence
No real request has completed through this route yet. Subnet 64 on taostats (opens in a new tab)
Documentation
Chutes documentation (opens in a new tab)

Image generation · Chutes

ADAPTER READY
Route id
chutes-image
Capability
Image
Subnet
Bittensor subnet 64 (Chutes)
Connection
Direct: POST https://vonkaiser-z-image-turbo.chutes.ai/generate · model Z-Image Turbo
Configuration
Server secret CHUTES_API_KEY. A Chutes API key (cpk_…) from chutes.ai/app/settings/api-keys, on a pay-as-you-go balance (Chutes has no free tier).
Cost
About $0.0005 per second of compute ($1.80 an hour), billed per started second. Paid by TAONIC; checked 2026-09-15.
Provenance
Chutes returns an invocation id with each call and publishes the chute’s record and instances. It does not publish which miner hotkey served a call, so no result is attributed to a particular miner.
Privacy
The question is sent to Chutes under its own terms.
Deadline
90 s
Evidence
No real request has completed through this route yet. Subnet 64 on taostats (opens in a new tab)
Documentation
Chutes documentation (opens in a new tab)

Text to speech · Chutes

ADAPTER READY
Route id
chutes-tts
Capability
Speech
Subnet
Bittensor subnet 64 (Chutes)
Connection
Direct: POST https://vonkaiser-audiodojo.chutes.ai/tts/kokoro · model Kokoro-82M
Configuration
Server secret CHUTES_API_KEY. A Chutes API key (cpk_…) from chutes.ai/app/settings/api-keys, on a pay-as-you-go balance (Chutes has no free tier).
Cost
About $0.0005 per second of compute ($1.80 an hour), billed per started second. Paid by TAONIC; checked 2026-09-15.
Provenance
Chutes returns an invocation id with each call and publishes the chute’s record and instances. It does not publish which miner hotkey served a call, so no result is attributed to a particular miner.
Privacy
The question is sent to Chutes under its own terms.
Deadline
60 s
Evidence
No real request has completed through this route yet. Subnet 64 on taostats (opens in a new tab)
Documentation
Chutes documentation (opens in a new tab)

Speech to text · Chutes

ADAPTER READY
Route id
chutes-stt
Capability
Speech
Subnet
Bittensor subnet 64 (Chutes)
Connection
Direct: POST https://vonkaiser-audiodojo.chutes.ai/stt/whisper · model Whisper large-v3-turbo
Configuration
Server secret CHUTES_API_KEY. A Chutes API key (cpk_…) from chutes.ai/app/settings/api-keys, on a pay-as-you-go balance (Chutes has no free tier).
Cost
About $0.0005 per second of compute ($1.80 an hour), billed per started second. Paid by TAONIC; checked 2026-09-15.
Provenance
Chutes returns an invocation id with each call and publishes the chute’s record and instances. It does not publish which miner hotkey served a call, so no result is attributed to a particular miner.
Privacy
The question is sent to Chutes under its own terms.
Deadline
90 s
Evidence
No real request has completed through this route yet. Subnet 64 on taostats (opens in a new tab)
Documentation
Chutes documentation (opens in a new tab)

Embeddings · Chutes

ADAPTER READY
Route id
chutes-embed
Capability
Find a match
Subnet
Bittensor subnet 64 (Chutes)
Connection
Direct: POST https://chutes-qwen-qwen3-embedding-8b-tee.chutes.ai/v1/embeddings · model Qwen/Qwen3-Embedding-8B-TEE
Configuration
Server secret CHUTES_API_KEY. A Chutes API key (cpk_…) from chutes.ai/app/settings/api-keys, on a pay-as-you-go balance (Chutes has no free tier).
Cost
About $0.0005 per second of compute ($1.80 an hour), billed per started second. Paid by TAONIC; checked 2026-09-15.
Provenance
Chutes returns an invocation id with each call and publishes the chute’s record and instances. It does not publish which miner hotkey served a call, so no result is attributed to a particular miner.
Privacy
The question is sent to Chutes under its own terms.
Deadline
30 s
Evidence
No real request has completed through this route yet. Subnet 64 on taostats (opens in a new tab)
Documentation
Chutes documentation (opens in a new tab)

Numinous, subnet 6

Numinous serves event forecasts made by miners on subnet 6, each attributed to the miner (uid and hotkey) and timed. Its public API needs no key.

Event forecasts · Numinous

VERIFIED
Route id
numinous-events
Capability
Market
Subnet
Bittensor subnet 6 (Numinous)
Connection
Direct: GET https://api.numinouslabs.io/api/v2/markets · model Numinous miner forecasts
Configuration
None: a keyless public API.
Cost
Free: public routes, no key. Numinous publishes no price list or rate limit. Paid by TAONIC; checked 2026-09-15.
Provenance
Numinous names the miner behind every forecast (uid and hotkey) and the time it was made; the event and the market price come from the same response.
Privacy
The question is sent to Numinous under its own terms.
Deadline
20 s
Evidence
Production request 30e55018-7fde-4f57-a7c6-b05056d7fe8d, 2026-09-24 05:16:47 UTC. Eight miners’ forecasts, each with its uid and hotkey, on Numinous event e673cd9f-8225-4dc4-8499-421779ac8057, answered through taonic.tech. Subnet 6 on taostats (opens in a new tab)
Documentation
Numinous documentation (opens in a new tab)

Desearch, subnet 22

Desearch serves live web and X search from subnet 22 miners, with sources.

ADAPTER READY
Route id
desearch-search
Capability
Search, Check a claim
Subnet
Bittensor subnet 22 (Desearch)
Connection
Direct: POST https://api.desearch.ai/desearch/ai/search/links/web · model Desearch web link search
Configuration
Server secret DESEARCH_API_KEY. A Desearch API key from console.desearch.ai ($5 of credit when you sign up and attach a card; no wallet). Desearch’s terms require its authorization to resell the service.
Cost
Usage-based, at the link-search rate Desearch shows in its Console; no public rate is published for this route. Public reference rates (7 May 2026): AI Search $0.40 and Web Search $1.00 per 1,000. Each answer reports its own cost, which TAONIC records. Paid by TAONIC; checked 2026-09-25.
Provenance
Every result carries its source URL; Desearch reports the cost of each call. It does not return which miner served the search, so no result is attributed to a particular miner.
Privacy
The question is sent to Desearch under its own terms.
Deadline
45 s
Evidence
No real request has completed through this route yet. Subnet 22 on taostats (opens in a new tab)
Documentation
Desearch documentation (opens in a new tab)

Synth, subnet 50

Synth serves probabilistic price forecasts from subnet 50 miners for a small set of assets.

Price forecasts · Synth

ADAPTER READY
Route id
synth-forecast
Capability
Market
Subnet
Bittensor subnet 50 (Synth)
Connection
Direct: GET https://api.synthdata.co/insights/prediction-percentiles · model Top-miner meta-model
Configuration
Server secret SYNTH_API_KEY. A paid Synth API key (from $49 a month) from synthdata.co; Synth’s written OK is needed to show its forecasts to TAONIC’s users.
Cost
Paid plans from $49 a month for 100 calls; 1 credit per call. Paid by TAONIC; checked 2026-09-15.
Provenance
Synth reports the forecast as a meta-model of its top-ranked miners, not per miner, so no forecast here is attributed to a particular miner.
Privacy
The question is sent to Synth under its own terms.
Deadline
15 s
Evidence
No real request has completed through this route yet. Subnet 50 on taostats (opens in a new tab)
Documentation
Synth documentation (opens in a new tab)

engy, subnet 53

engy serves text generation on subnet 53. Its own terms say most models run on engy’s GPUs and only some reach public miners. It is a fallback route for text in TAONIC’s registry, and has never been connected.

Text generation · engy

ADAPTER READY
Route id
engy-chat
Capability
Answer, Check a claim
Subnet
Bittensor subnet 53 (engy)
Connection
Direct: POST https://api.engy.ai/v1/chat/completions · model qwen3.6-35b-a3b
Configuration
Server secret ENGY_API_KEY. An engy API key from an engy.ai account (email or Google; no wallet), on a paid balance.
Cost
$0.045 per million input tokens, $0.30 per million output tokens and $0.015 per million cached input tokens, prepaid. Paid by TAONIC; checked 2026-09-25.
Provenance
engy returns a completion id, and TAONIC keeps an x-engy-request-id header when one is sent; engy does not return which miner served the request, so no answer is attributed to a particular miner.
Privacy
The question is sent to engy under its own terms.
Deadline
60 s
Evidence
No real request has completed through this route yet. Subnet 53 on taostats (opens in a new tab)
Documentation
engy documentation (opens in a new tab)

Energon and engy

engy is the subnet 53 service. Energon is an independent, pre-launch front end for engy with its own token and its own reserve; it is not part of TAONIC and TAONIC is not part of it. The only relationship that exists is the one above: an engy adapter in TAONIC’s registry that has never been configured or connected.

What an integration could look like

Possible integration layers with engy, and their state
LayerState
TAONIC routes compatible text work to engy on subnet 53ADAPTER READY
TAONIC records the provider cost and route evidence for each such requestDESIGN
engy is paid for completed work, as any provider isDESIGN
A provider-growth allocation, only under an explicit written agreement that defines custody, ownership, accounting and withdrawal rightsUNAVAILABLE

Proposing a route

There is no self-serve onboarding. A subnet service can be routed to when it serves a documented, reachable API that its miners answer, publishes terms that allow resale or display, and returns enough provenance to attribute an answer. Proposals go to contact@taonic.tech.

Every state on this page matches Status. Where they ever differ, Status is right and this page is a defect.