io.github.InferIndex/inferindex

Est-ce que io.github.InferIndex/inferindex est sûr ?

Au dernier audit MCP Hub, le serveur MCP « io.github.InferIndex/inferindex » est classé « Non encore testé ». Référencé dans l'annuaire. Seul le pentest comportemental fait foi, et il n'a pas encore été exécuté sur ce serveur. Le verdict est réévalué en continu ; consultez la fiche pour l'état à jour.

Méthode : MCP Hub analyse la description de chaque outil (couche 1) et, pour le verdict fort, exécute le serveur dans une micro-VM isolée avec leurres et réseau bloqué (couche 2). On publie des noms et des compteurs, jamais le contenu brut renvoyé par un outil.

Les 8 outils

• cheapest≈ 770 tokens

  Cheapest current API offers for one model across direct providers and aggregators, in USD per 1M tokens (input, output, blended 3:1). Stale prices, and flex/batch tiers, are excluded by default. Optional filters (context, tools, JSON, vision, region, no training on prompts, open sign-up) and usage (tokens per request, requests per day) to get an estimated cost per request and per month. Returns the winner in detail and one short line per following offer. A condition that is absent was not published by the provider, it never means "no"; a flag not listed in signals is false.

  Paramètres, retour & exemple (16)

  Paramètres

  • json · boolean · optionnel
    Only offers that support JSON output
  • limit · integer · optionnel
    Number of offers to return, winner included (default 5, max 25)
  • model · string · requis
    Model id or name, e.g. 'deepseek-v3.2', 'deepseek/deepseek-v4-pro', 'gpt-5.6-luna'. Use search_models when unsure.
  • tools · boolean · optionnel
    Only offers that support tool calling
  • detail · string · optionnel · compact | full
    compact (default): the winner in detail and one short line per other offer (provider, prices, signals, training on prompts when published). full: every field of every offer (all published conditions, reliability), much longer
  • region · string · optionnel
    Only providers that process data in this region: eu, us, …
  • strict · boolean · optionnel
    Exclude offers whose provider does not publish the filtered information (by default they are kept and flagged)
  • vision · boolean · optionnel
    Only offers that accept image input
  • min_context · integer · optionnel
    Minimum context window in tokens
  • no_training · boolean · optionnel
    Only providers whose published terms say they do not train on your prompts
  • no_waitlist · boolean · optionnel
    Only providers with open sign-up (no waitlist, invitation or country restriction)
  • cached_ratio · number · optionnel
    Share of input tokens served from the provider's prompt cache (0 to 1)
  • include_tiers · string · optionnel
    Also include lower-priority service tiers, comma-separated: flex, batch (hidden by default)
  • output_tokens · integer · optionnel
    Output tokens per request, for the estimated cost
  • prompt_tokens · integer · optionnel
    Input tokens per request, for the estimated cost
  • requests_per_day · integer · optionnel
    Requests per day, to also get an estimated monthly cost

  Exemple d'appel

  {"name": "cheapest", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• compare_providers≈ 700 tokens

  Current offers for one model, one line per provider and source (direct or via an aggregator), cheapest first (10 by default), with price, quantization, signals and training on prompts when published; the cheapest offer comes in detail. Pass detail: "full" for context, every published condition (data regions, sign-up) and reliability from official status pages on each offer.

  Paramètres, retour & exemple (13)

  Paramètres

  • sort · string · optionnel · blended | input | output | estimated_cost
    Sort order: default blended (3:1 mix), or estimated_cost (the cost of your load) when prompt_tokens or output_tokens is given; estimated_cost needs one of them
  • limit · integer · optionnel
    Number of offers to return, cheapest included (default 10, max 50)
  • model · string · requis
    Model id or name, e.g. 'deepseek-v3.2', 'deepseek/deepseek-v4-pro', 'gpt-5.6-luna'. Use search_models when unsure.
  • detail · string · optionnel · compact | full
    compact (default): the winner in detail and one short line per other offer (provider, prices, signals, training on prompts when published). full: every field of every offer (all published conditions, reliability), much longer
  • region · string · optionnel
    Only providers that process data in this region: eu, us, …
  • strict · boolean · optionnel
    Exclude offers whose provider does not publish the filtered information (by default they are kept and flagged)
  • no_training · boolean · optionnel
    Only providers whose published terms say they do not train on your prompts
  • no_waitlist · boolean · optionnel
    Only providers with open sign-up (no waitlist, invitation or country restriction)
  • cached_ratio · number · optionnel
    Share of input tokens served from the provider's prompt cache (0 to 1)
  • include_tiers · string · optionnel
    Also include lower-priority service tiers, comma-separated: flex, batch (hidden by default)
  • output_tokens · integer · optionnel
    Output tokens per request, for the estimated cost
  • prompt_tokens · integer · optionnel
    Input tokens per request, for the estimated cost
  • requests_per_day · integer · optionnel
    Requests per day, to also get an estimated monthly cost

  Exemple d'appel

  {"name": "compare_providers", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• estimate_cost≈ 450 tokens

  Estimated cost of a workload on one model at each provider: cost per request, and per month if requests_per_day is given, taking the provider's tiered pricing and prompt-cache price into account. Offers sorted by estimated cost, cheapest first: the cheapest in detail, then one short line per offer with its estimated cost.

  Paramètres, retour & exemple (7)

  Paramètres

  • limit · integer · optionnel
    Number of offers to return, cheapest included (default 5, max 25)
  • model · string · requis
    Model id or name, e.g. 'deepseek-v3.2', 'deepseek/deepseek-v4-pro', 'gpt-5.6-luna'. Use search_models when unsure.
  • detail · string · optionnel · compact | full
    compact (default): the winner in detail and one short line per other offer (provider, prices, signals, training on prompts when published). full: every field of every offer (all published conditions, reliability), much longer
  • cached_ratio · number · optionnel
    Share of input tokens served from the provider's prompt cache (0 to 1)
  • output_tokens · integer · optionnel
    Output tokens per request, for the estimated cost
  • prompt_tokens · integer · optionnel
    Input tokens per request, for the estimated cost
  • requests_per_day · integer · optionnel
    Requests per day, to also get an estimated monthly cost

  Exemple d'appel

  {"name": "estimate_cost", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• gpu_rentals≈ 399 tokens

  Rental offers for one GPU type: provider, price in USD per GPU per hour, number of GPUs in the published configuration, billing and region when published. One block per tier (guaranteed, community, spot), each with its own cheapest offer, also given per configuration (cheapest_by_gpu_count: the price per GPU of a single GPU and of a multi-GPU node are not comparable); tiers are never compared with each other. guaranteed: capacity the provider presents as not interrupted; community: third-party hosts; spot: interruptible capacity. Prices are as published by the provider, and dated. Get the GPU id from list_gpus.

  Paramètres, retour & exemple (5)

  Paramètres

  • gpu · string · requis
    GPU id from list_gpus, for example h100-sxm-80gb
  • tier · string · optionnel · guaranteed | community | spot
    Only this tier (default: one block per tier)
  • detail · string · optionnel · compact | full
    compact (default): the 3 cheapest offers of each tier. full: up to 50 offers per tier
  • region · string · optionnel
    Only this region, as published by the provider
  • gpu_count · integer · optionnel
    Only this published configuration: 1, 2, 4, 8 or 16 GPUs

  Exemple d'appel

  {"name": "gpu_rentals", "arguments": {"gpu": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• list_gpus≈ 169 tokens

  GPU types whose rental price InferIndex tracks (H100, A100, L40S…), with the lowest price per GPU per hour in USD for each tier (guaranteed, community, spot) and the number of GPUs of that configuration. Prices are per GPU per hour, as published by each provider, and dated. Tiers are different products and are never compared with each other.

  Paramètres, retour & exemple (0)

  Exemple d'appel

  {"name": "list_gpus", "arguments": {}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• price_history≈ 364 tokens

  Price history of one model: every offer tracked by InferIndex (daily or weekly min/max/last price in USD, or raw price changes), plus the official price of the model's lab over time. Give either days, or from/to (YYYY-MM-DD), or at (a date) for the prices in effect that day.

  Paramètres, retour & exemple (8)

  Paramètres

  • at · string · optionnel
    A single date, YYYY-MM-DD: prices in effect that day
  • to · string · optionnel
    End date, YYYY-MM-DD
  • days · integer · optionnel
    Number of days back from today (default 7)
  • from · string · optionnel
    Start date, YYYY-MM-DD (with to, instead of days)
  • limit · integer · optionnel
    Maximum number of points (default 100, max 500)
  • model · string · requis
    Model id or name, e.g. 'deepseek-v3.2', 'deepseek/deepseek-v4-pro', 'gpt-5.6-luna'. Use search_models when unsure.
  • provider · string · optionnel
    Only this provider
  • granularity · string · optionnel · day | week | raw
    day (default), week, or raw price changes

  Exemple d'appel

  {"name": "price_history", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• search_models≈ 142 tokens

  Find the exact id of an LLM tracked by InferIndex from a name or partial name (e.g. 'deepseek', 'qwen3 max', 'claude opus'). Returns matching model ids and names, best match first.

  Paramètres, retour & exemple (1)

  Paramètres

  • query · string · requis
    Model name or part of it

  Exemple d'appel

  {"name": "search_models", "arguments": {"query": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• self_host_or_api≈ 1011 tokens

  Is it cheaper to host an open-weights model yourself on rented GPUs, or to use the cheapest API offer? Returns a one-sentence verdict (headline) and the numbers behind it: the break-even volume in million tokens per day, the throughput the whole configuration must deliver for self-hosting to cost less (required_tokens_per_second: compare it with what your setup does), the break-even utilization when a throughput is known, the GPU configuration and its hourly price, and what the verdict rests on (confidence). Without a volume you still get the verdict and the break-even volume. It is an estimate: the throughput is a published measurement (measured), an estimate scaled from one (derived; estimated for a wide range), or a published minimum (lower_bound: a floor, valid for requests of up to 2,048 tokens in total, that can show self-hosting wins, never that the API is cheaper); or you give your own tokens_per_second. When no published figure decides, the headline gives the break-even volume and the required throughput, to compare with yours, rather than a yes or no, and may point out a dearer setup that published figures do decide (settled_alternative: for comparison, not a recommendation). headline_kind and headline_values give the headline as fields. It declines to give a number (status refused, with a reason) when it cannot do so reliably, for example a closed model, weights that do not fit the requested GPU, or a GPU nobody rents. Hardware rental only: the hourly price the provider publishes for the machine. It does not count the engineering to deploy and run the model, monitoring, redundancy, start-up time, or storage and network costs billed separately.

  Paramètres, retour & exemple (13)

  Paramètres

  • gpu · string · optionnel
    Impose a GPU: id from list_gpus, for example h100-sxm-80gb
  • tier · string · optionnel · guaranteed | community | spot
    GPU rental tier (default guaranteed)
  • model · string · requis
    Model id or name, e.g. 'deepseek-v3.2', 'deepseek/deepseek-v4-pro', 'gpt-5.6-luna'. Use search_models when unsure.
  • detail · string · optionnel · compact | full
    compact (default): verdict, headline and key numbers. full: also the three throughput scenarios, the model and every assumption
  • gpu_count · integer · optionnel
    Impose a number of GPUs: 1, 2, 4, 8 or 16
  • utilization · number · optionnel
    Average share of time the GPUs serve requests, in percent (default 50)
  • quantization · string · optionnel · fp16 | bf16 | fp8 | int8 | fp4 | int4
    Precision of the weights (default fp8); must be published by at least one provider unless fp16 or bf16
  • output_tokens · integer · optionnel
    Output tokens per request
  • prompt_tokens · integer · optionnel
    Input tokens per request
  • tokens_per_day · integer · optionnel
    Your volume in tokens per day (input and output together). Or give requests_per_day, prompt_tokens and output_tokens
  • weights_margin · integer · optionnel
    Percent of the GPU memory the model weights may fill (default 70; the rest is the engine reserve and the context cache)
  • requests_per_day · integer · optionnel
    Requests per day (with prompt_tokens and output_tokens)
  • tokens_per_second · number · optionnel
    Your own measured throughput for the whole configuration, input and output tokens together, instead of our estimate

  Exemple d'appel

  {"name": "self_host_or_api", "arguments": {"model": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

C'est votre serveur ?

Revendiquez-le sur l'espace créateur (preuve GitHub, domaine ou npm) pour afficher le badge « propriétaire vérifié » et lancer un pentest comportemental à la demande afin d'obtenir le badge « Vérifié ».

Créer un compte / revendiquer →

Source

https://mcp.inferindex.dev/mcp