AI Horde
Community-supplied inference across 26 models; capacity varies and throughput is not guaranteed.
https://aihorde.net/api/v2Models mentioned
aphrodite/SicariusSicariiStuff/Impish_LLAMA_4Baphrodite/TheDrummer/Behemoth-X-123B-v2.1aphrodite/TheDrummer/Behemoth-128B-v3-NVFP4aphrodite/TheDrummer/Cydonia-24B-v4.3aphrodite/TheDrummer/Skyfall-31B-v4.2koboldcpp/Angelic_Eclipse-12Bkoboldcpp/Judas-Uncensored-3.2-1B.Q8koboldcpp/L3-8B-Stheno-v3.2-Q5_K_Mkoboldcpp/gemma-4-31B-it-heretickoboldcpp/gemma-4-E2B-it-qat-UD-Q4_K_XL.ggufkoboldcpp/Gemma-4-E4B-it-Ultra-Uncensored-Heretickoboldcpp/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q4_K_Mkoboldcpp/L3-8B-Stheno-v3.2koboldcpp/L3-8B-Stheno-v3.2-Q8_0koboldcpp/Llama-3.2-1B-Instructkoboldcpp/Llama-3.2-3Bkoboldcpp/Llama-3.2-3B-Instruct-Q4_K_Mkoboldcpp/Qwen2.5-1.5Bkoboldcpp/Qwen3-4B-Nymphaea-RP.i1-Q4_K_Skoboldcpp/Qwen3.8-27B-Uncensoredkoboldcpp/Qwen_Qwen3-0.6B-IQ4_XSkoboldcpp/Qwen/Qwen3.5-0.8Bkoboldcpp/ReadyArt/Forgotten-Safeword-22Bkoboldcpp/debug-L3-8B-Stheno-v3.2koboldcpp/mini-magnum-12b-v1.1meta-llama/Llama-3.2-3B-InstructEquivalent paid value
How this was valued: The published allowance or a defensible paid comparison is not precise enough to calculate.
Capacity is donated and prioritized with non-purchasable Kudos rather than a fixed monetary or token allowance.
Limits and terms
- Mechanism
- Nonmonetary kudos and queue priority rather than a fixed request quota.
- Anonymous Priority
- lowest
- Load Shedding
- Anonymous use can be restricted during load.
- Monetary Purchase
- Kudos cannot be bought or sold.
What happens to your prompts?
The coordinator is transient, but volunteer workers can inspect or retain plaintext requests outside a uniform contractual control.
- Plan Scope
- AI Horde's central queue plus independently operated volunteer workers for text and image generation.
- Prompt Retention
- The central Horde says prompts and generations are held transiently in memory and deleted shortly after delivery or cancellation. A worker operator can technically modify open-source worker software to view and save every prompt.
- Response Retention
- Same central transient retention, with no enforceable guarantee that a volunteer worker did not save its generated result.
- Ordinary Logging
- The service stores account identifiers and operational queue, kudos, request, worker, and safety data; the public FAQ does not publish a single retention schedule for all metadata.
- Model Training
- The central service says it does not store prompt or generation details, but there is no contractual mechanism preventing a volunteer worker from saving or reusing received content, including for training.
- Product Improvement
- No central content-based improvement program is disclosed; safety filters inspect requests and workers can independently modify their software.
- Human Or Operator Access
- Each selected worker receives the full prompt and generates the output. Official documentation explicitly says technically capable worker operators can spy on and save both.
- Subprocessors And Routing
- Requests are distributed to community-run computers outside AI Horde's direct physical or contractual control; workers do not receive the requester's ID or IP but do receive content.
- Deletion Controls
- Central prompt and generation state is deleted shortly after completion or cancellation. No central deletion control can retract data a worker operator may have copied.
- Caveat
- Treat every request as if posting to a public forum. The coordinator's transient storage does not make an untrusted volunteer-compute network suitable for confidential or personal data.
Governing documents
- Terms Of Service
- https://aihorde.net/terms/
- Privacy Policy
- https://aihorde.net/privacy
Eligibility
- Account Required
- No
- Payment Method Required
- No
Primary sources
Before you build with AI Horde
Read the classification narrowly
This record has current first-party evidence for some zero-cost hosted inference. The label describes the bounded offer supported by evidence on 2026-09-19; it does not guarantee permanence, production suitability, uptime, latency, model quality, or access from every region.
Resolve the live model route
This snapshot records 26 model IDs. Match the exact ID against the provider's live catalog before using it in code; zero-price routes and aliases can rotate while an older documentation page remains online. The recorded base endpoint is https://aihorde.net/api/v2, but the provider's current API reference remains authoritative for paths, authentication, and request shape.
Confirm account and billing boundaries
The recorded offer does not require an account, but published limits and abuse controls can still change without notice. The reviewed evidence explicitly says no payment card is required. Re-check the signup flow because eligibility and billing controls can change after the snapshot. The equivalent paid value shown here prices a documented allowance where possible; it is not cash, guaranteed savings, or protection from overage.
Re-read the prompt policy for your plan
This record classifies the reviewed handling as “Partially private”: The coordinator is transient, but volunteer workers can inspect or retain plaintext requests outside a uniform contractual control. Provider policies can distinguish free, paid, enterprise, opted-in, and feature-specific traffic, so verify the governing terms for the exact account and route that will receive your data.
Follow the evidence, then re-check it
The record links 8 first-party sources covering the offer, catalog, limits, pricing, terms, privacy, or adoption evidence available to the audit. Prefer the newest governing document or live catalog when sources disagree, and submit a correction when a provider changes a material term.
Match the quota to the workload shape
Translate the published allowance into the traffic pattern you actually expect instead of comparing headline totals alone. A daily token pool can look generous while a low requests-per-minute or concurrency ceiling blocks interactive bursts; a high request limit can still fail a long-context job when input, output, or per-request tokens are capped. Separate prompt tokens from generated tokens, include retries and tool calls, and test the largest realistic payload. If the provider documents more than one limit window, the tightest window at your peak load is the practical ceiling. Research, experimental, and community access can also carry eligibility or fair-use constraints that cannot be modeled as a simple number.
Plan fallback without changing the rules
A fallback should preserve more than API syntax. Confirm that the substitute route supports the required modality, context length, streaming behavior, structured output, tools, and safety controls, then compare its prompt-retention and training terms. An OpenAI-compatible request shape does not make providers operationally or contractually equivalent. Decide which errors may trigger a retry, cap retry storms, and prevent an exhausted free route from silently switching to a billable model. If deterministic output matters, record the model revision and sampling settings; rotating aliases and free-model pools can change behavior even when the endpoint remains available.
Monitor the offer as a dependency
Capture the model ID, response model field, rate-limit headers, usage fields, latency, HTTP status, and any provider request identifier for each test. Watch for authorization failures, quota exhaustion, catalog removal, policy revisions, and dashboard balance changes as separate failure modes. Re-check the provider’s live catalog and governing pages on a schedule proportionate to the workload’s importance, and keep the dated sources that supported your decision. Free capacity is especially suitable for prototypes, evaluations, fallbacks, and bounded workloads when the application can tolerate change; a production dependency still needs observability, an exit path, and an owner responsible for re-verification.
Test one complete request before scaling
Start with the smallest permitted request using the exact credential, model ID, endpoint, and account type you intend to deploy. Record the HTTP status, response headers, usage fields, latency, and any dashboard balance change. Then exercise the failure path: an invalid model, an exhausted quota, or a rate-limit response should fail clearly without silently switching to a paid route. If the provider supports streaming or tool calls, validate those features separately because a free model can expose a narrower capability set than its paid counterpart. Keep a budget ceiling outside the application whenever billing is possible, and avoid sending sensitive data until the observed route matches the reviewed data agreement.