robots.txt
# General Compute — robots.txt
# We are agent-forward. AI crawlers, search crawlers, and autonomous agents are welcome.
# Sign-up and full API docs for agents: https://docs.generalcompute.com
User-agent: *
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
Disallow: /sso-callback
# ---------- OpenAI ----------
User-agent: GPTBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: OAI-SearchBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: ChatGPT-User
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Anthropic ----------
User-agent: ClaudeBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Claude-Web
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Claude-User
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Claude-SearchBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: anthropic-ai
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Google ----------
User-agent: Google-Extended
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: GoogleOther
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Perplexity ----------
User-agent: PerplexityBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Perplexity-User
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Meta ----------
User-agent: Meta-ExternalAgent
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Meta-ExternalFetcher
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Apple ----------
User-agent: Applebot-Extended
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
# ---------- Other AI crawlers ----------
User-agent: Bytespider
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: CCBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: cohere-ai
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: cohere-training-data-crawler
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Diffbot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: DuckAssistBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: FriendlyCrawler
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: ImagesiftBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Kagibot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: MistralAI-User
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: Timpibot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
User-agent: YouBot
Content-Signal: search=yes, ai-train=yes, ai-input=yes
Allow: /
Disallow: /api/
Sitemap: https://www.generalcompute.com/sitemap.xml
llms.txt
# General Compute
> We deploy SambaNova, Cerebras, Positron and d-Matrix. Prefill
> on GPUs, decode on purpose-built silicon; dedicated racks under one contract.
> OpenAI-compatible API.
Developer resources for General Compute live on this domain: API documentation,
OpenAPI specification, authentication, webhooks, MCP server, and SDKs. Every
page is also available as markdown: request it with `Accept: text/markdown` or
append `.md` to the path. A full-text dump lives at https://www.generalcompute.com/llms-full.txt.
## Key facts
- API base URL: `https://api.generalcompute.com/v1`
- Protocol: OpenAI-compatible HTTPS
- Agent signup (no human-only gate): https://docs.generalcompute.com/agent-signup
- Docs: https://docs.generalcompute.com
- Contact: jason@generalcompute.com
## Developer resources
- [General Compute products](https://www.generalcompute.com/products)
- [General Compute pricing](https://www.generalcompute.com/pricing)
- [General Compute API documentation](https://docs.generalcompute.com)
- [General Compute API reference](https://www.generalcompute.com/api-reference)
- [General Compute OpenAPI specification](https://docs.generalcompute.com/openapi.json)
- [General Compute authentication & API keys](https://www.generalcompute.com/auth)
- [General Compute webhooks](https://www.generalcompute.com/webhooks)
- [General Compute MCP server](https://www.generalcompute.com/mcp)
- [General Compute SDKs](https://www.generalcompute.com/sdks)
- [General Compute agent signup](https://docs.generalcompute.com/agent-signup)
- [OpenCode integration](https://www.generalcompute.com/opencode)
- [General Compute MCP descriptor](https://www.generalcompute.com/.well-known/mcp)
- [General Compute API catalog (RFC 9727)](https://www.generalcompute.com/.well-known/api-catalog)
- [General Compute agent skills](https://www.generalcompute.com/.well-known/agent-skills/index.json)
## Pages
- [Home](https://www.generalcompute.com/.md): What General Compute is: the neocloud for SambaNova, Cerebras, Positron and d-Matrix. Prefill on GPUs, decode on purpose-built silicon, plus a quick-start snippet.
- [Products](https://www.generalcompute.com/products.md): General Compute API access, custom deployments, and bring-your-own-model serving.
- [Pricing](https://www.generalcompute.com/pricing.md): Self-serve API credit, dedicated capacity, and private model deployment options.
- [Agents portal](https://www.generalcompute.com/agents.md): Entry point for automated consumers — discovery resources and machine rules.
- [General Compute developer resources](https://www.generalcompute.com/developers.md): Hub for General Compute API docs, OpenAPI spec, authentication, webhooks, MCP server, and SDKs.
- [General Compute API reference](https://www.generalcompute.com/api-reference.md): OpenAI-compatible API base URL, models, auth header, and a curl example for General Compute.
- [General Compute OpenAPI specification](https://www.generalcompute.com/openapi.md): Where to fetch the General Compute OpenAPI JSON and how to consume it.
- [General Compute authentication](https://www.generalcompute.com/auth.md): Bearer-token auth, key rotation, and the agent signup flow for General Compute.
- [General Compute webhooks](https://www.generalcompute.com/webhooks.md): Webhook event categories, HMAC signing, and retry behaviour for General Compute.
- [General Compute MCP server](https://www.generalcompute.com/mcp.md): Model Context Protocol endpoint and discovery descriptor for the General Compute MCP server.
- [General Compute SDKs](https://www.generalcompute.com/sdks.md): Python and Node clients for General Compute, with install and usage snippets.
- [OpenClaw integration](https://www.generalcompute.com/openclaw.md): Swap an OpenAI-compatible coding agent to General Compute.
- [OpenCode integration](https://www.generalcompute.com/opencode.md): Configure OpenCode with General Compute and use $100 free credit for coding-agent workloads.
- [Benchmarks](https://www.generalcompute.com/benchmarks.md): Head-to-head latency and throughput vs other inference providers.
- [Inference is fragmenting](https://www.generalcompute.com/whitepaper.md): General Compute whitepaper on why inference is fragmenting, the specialized silicon winning each region of the Pareto curve, and the balance sheet that deploys it.
- [Infrastructure](https://www.generalcompute.com/infrastructure.md): Purpose-built ASIC clusters, energy economics, and architecture.
- [Coding agents](https://www.generalcompute.com/use-cases/coding-agents.md): Workload profile and recommended settings for autonomous coding agents.
- [Voice AI](https://www.generalcompute.com/use-cases/voice-ai.md): Latency profile and streaming notes for real-time voice workloads.
- [Roadmap](https://www.generalcompute.com/roadmap.md): Current site, expansion plan, and deployment playbook.
- [Team](https://www.generalcompute.com/team.md): Operators to contact for rate-limit changes or escalation.
- [Demo](https://www.generalcompute.com/demo.md): Pointer to the interactive demo and programmatic benchmarking.
- [Blog](https://www.generalcompute.com/blog.md): Posts on inference performance, model serving, and agent infrastructure.
- [Announcements](https://www.generalcompute.com/announcements.md): Company announcements from General Compute.
- [Terms of service](https://www.generalcompute.com/terms.md): Agent-readable summary of the platform terms.
- [Privacy](https://www.generalcompute.com/privacy.md): Agent-readable summary of the privacy policy.
## Blog
- [We Signed a Multi-Year Agreement to Deploy Cerebras Wafer-Scale Inference](https://www.generalcompute.com/blog/general-compute-cerebras-multi-year-agreement.md): General Compute will bring Cerebras wafer-scale inference to its customers at scale. Capacity opens in Q1 2027, and agentic coding is the first workload. Here is what we are doing and why.
- [How Transformer Architecture Determines Inference Speed and Memory Usage](https://www.generalcompute.com/blog/how-transformer-architecture-determines-inference-speed-and-memory-usage.md): A deep look at how the architectural decisions baked into a transformer model at training time -- attention variant, layer count, hidden size, and more -- directly determine how fast and how cheaply it can run at inference time.
- [What Is an AI Inference Server? Cloud API vs Self-Hosted vs On-Premise Explained](https://www.generalcompute.com/blog/what-is-an-ai-inference-server.md): An AI inference server accepts prompts and returns model completions. Here's how they work, the main deployment options, and how to choose between managed APIs and self-hosted solutions.
- [Why TileRT Cannot Catch SambaNova SN50: The Limits of Software-Emulated Dataflow on the Wrong Hardware](https://www.generalcompute.com/blog/why-tilert-cannot-catch-sambanova-sn50.md): TileRT is a clever software approach to dataflow execution, but it runs on hardware that was never designed for reconfigurable dataflow. Here's why purpose-built silicon like the SambaNova SN50 will always hold the edge.
- [AI Inference Costs in 2025: How to Calculate and Reduce Your LLM API Spend](https://www.generalcompute.com/blog/ai-inference-costs-2025-how-to-calculate-and-reduce-your-llm-api-spend.md): A practical guide to understanding LLM API pricing, calculating your monthly inference bill with real formulas, and eight proven strategies to cut costs without sacrificing quality.
- [Fast AI Inference: 8 Proven Techniques That Deliver Real-World Speedups](https://www.generalcompute.com/blog/fast-ai-inference.md): A technical guide to the eight most effective LLM inference optimization techniques, with concrete speedup data, code examples, and stacking strategies for production deployments.
- [What Is Tokens Per Second (TPS)? The LLM Throughput Metric That Actually Matters](https://www.generalcompute.com/blog/what-is-tokens-per-second-tps.md): Tokens per second (TPS) measures how fast a language model generates output. Here's what it means, how it differs from TTFT, what drives it, and how to benchmark it in your own stack.
- [What Is Time to First Token (TTFT)? The LLM Latency Metric That Defines UX](https://www.generalcompute.com/blog/what-is-time-to-first-token-ttft.md): Time to First Token (TTFT) measures how long it takes an LLM to generate its first output token after receiving your request. It's the metric that determines whether your AI app feels instant or sluggish.
- [LLM Inference Explained: How It Works, What It Costs, and How to Optimize It](https://www.generalcompute.com/blog/llm-inference-explained.md): A complete technical guide to LLM inference: the prefill/decode pipeline, the metrics that matter, how providers price token generation, and five concrete ways to make it faster and cheaper.
- [The 1B-and-Under Club: TinyLlama vs Phi-2 vs StableLM vs Qwen 0.5B](https://www.generalcompute.com/blog/the-1b-and-under-club-tinyllama-vs-phi-2-vs-stablelm-vs-qwen-0-5b.md): A benchmark comparison of the smallest capable language models -- TinyLlama 1.1B, Phi-2 2.7B, StableLM 2 1.6B, and Qwen 2.5 0.5B -- covering reasoning, coding, inference speed, memory footprint, and which to use for edge and mobile deployments.
- [Time-to-First-Token Leaderboard: Every Major Open Model Ranked](https://www.generalcompute.com/blog/time-to-first-token-leaderboard-every-major-open-model-ranked.md): TTFT benchmarks across every major open-source model at multiple prompt lengths. We measured prefill latency on identical hardware so you can pick the right model for latency-sensitive workloads.
- [Batch Inference Throughput: Which Models Maximize Tokens-per-Dollar?](https://www.generalcompute.com/blog/batch-inference-throughput-which-models-maximize-tokens-per-dollar.md): A throughput-focused benchmark comparing open-source models across batch sizes. We measured tokens-per-second, cost-per-million-tokens, and effective tokens-per-dollar so you can pick the right model for high-volume workloads.
- [Long-Context Benchmarks: Who's Actually Fast at 128K Tokens?](https://www.generalcompute.com/blog/long-context-benchmarks-128k-tokens.md): We measured TTFT, decode speed, and effective throughput for major open-source models at 128K token context. Here's what the numbers show.
- [MoE vs Dense at Equal Compute: Mixtral 8x7B vs Llama 70B](https://www.generalcompute.com/blog/moe-vs-dense-equal-compute-mixtral-8x7b-vs-llama-70b.md): Mixtral 8x7B activates ~13B parameters per token while Llama 70B activates 70B. At equal active FLOPs, MoE architecture delivers dramatically better quality than dense. Here's why, with concrete benchmarks.
- [Quantized vs Full Precision: GPTQ, AWQ, and FP8 Benchmarks Across Model Sizes](https://www.generalcompute.com/blog/quantized-vs-full-precision-gptq-awq-fp8-benchmarks-across-model-sizes.md): A practical comparison of GPTQ, AWQ, and FP8 quantization against FP16 full precision across 7B, 13B, and 70B model sizes, covering throughput, latency, memory, and quality impact.
- [Instruction-Tuned vs Base Models: Does Fine-Tuning Cost You Speed?](https://www.generalcompute.com/blog/instruction-tuned-vs-base-models-does-fine-tuning-cost-you-speed.md): A practical look at whether instruction tuning and RLHF actually affect inference speed, what the real sources of performance differences are, and how to benchmark fairly across checkpoints.
- [Vision-Language Models: Llama 4 Scout vs Qwen2-VL vs InternVL at Inference](https://www.generalcompute.com/blog/vision-language-models-llama-4-scout-vs-qwen-vl-vs-internvl-at-inference.md): A practical comparison of Llama 4 Scout, Qwen2-VL, and InternVL2 on multimodal benchmarks and inference speed, with guidance on which model fits which deployment scenario.
- [Embedding Model Benchmarks: Speed vs Quality for RAG Retrieval](https://www.generalcompute.com/blog/embedding-model-benchmarks-speed-vs-quality-for-rag-retrieval.md): A benchmark comparison of embedding models on retrieval quality (BEIR NDCG@10) and inference latency, with practical guidance on where the quality tradeoffs actually matter for RAG pipelines.
- [Math and Reasoning Models: DeepSeek-R1 vs Qwen QwQ vs Llama Reasoning](https://www.generalcompute.com/blog/math-and-reasoning-models-deepseek-r1-vs-qwen-qwq-vs-llama-reasoning.md): A benchmark comparison of the leading open-source thinking/reasoning models -- DeepSeek-R1, QwQ-32B, and Llama-based reasoning distillations -- covering accuracy, token costs, and when to use each.
- [Coding Model Benchmarks: Qwen 2.5 Coder vs DeepSeek Coder V2 vs CodeLlama vs StarCoder 2](https://www.generalcompute.com/blog/coding-model-benchmarks-qwen-2-5-coder-vs-deepseek-coder-v2-vs-codellama-vs-starcoder-2.md): A technical benchmark comparison of the top open-source coding models -- Qwen 2.5 Coder, DeepSeek Coder V2, CodeLlama, and StarCoder 2 -- across HumanEval, MBPP, and real-world code tasks by size bucket.
- [Frontier Open Models: Llama 4 Maverick vs DeepSeek V3 vs Qwen 2.5 110B](https://www.generalcompute.com/blog/frontier-open-models-llama-4-maverick-vs-deepseek-v3-vs-qwen-2-5-110b.md): A technical comparison of the top frontier open-source models -- Llama 4 Maverick, DeepSeek V3, and Qwen 2.5 110B -- covering architecture, benchmarks, and inference trade-offs.
- [Mid-Size Models: Llama 3.1 70B vs Qwen 2.5 72B vs Mistral Large vs DeepSeek-V2 Lite](https://www.generalcompute.com/blog/mid-size-models-llama-3-1-70b-vs-qwen-2-5-72b-vs-mistral-large-vs-deepseek-v2-lite.md): A benchmark comparison of the 70B-class open models in 2025 -- covering reasoning, math, coding, context length, and inference speed across Llama 3.1 70B, Qwen 2.5 72B, Mistral Large 2, and DeepSeek-V2 Lite.
- [The 7-8B Battle: Llama 3.1 8B vs Mistral 7B v0.3 vs Qwen 2.5 7B vs Gemma 2 9B](https://www.generalcompute.com/blog/the-7-8b-battle-llama-3-1-8b-vs-mistral-7b-vs-qwen-2-5-7b-vs-gemma-2-9b.md): A benchmark comparison of the four most popular 7-9B parameter models in 2025 -- covering reasoning, math, coding, context length, and inference speed to help you pick the right model for your application.
- [Small Models Showdown: Qwen 2.5 3B vs Llama 3.2 3B vs Phi-3.5 Mini vs Gemma 2 2B](https://www.generalcompute.com/blog/small-models-showdown-qwen-2-5-3b-llama-3-2-3b-phi-3-5-mini-gemma-2-2b.md): A head-to-head benchmark comparison of the best sub-4B parameter models in 2025 -- covering reasoning, coding, math, context length, and inference speed to help you pick the right small model for your use case.
- [Agentic RAG: When Retrieval and Generation Both Need to Be Fast](https://www.generalcompute.com/blog/agentic-rag-when-retrieval-and-generation-both-need-to-be-fast.md): In agentic RAG, retrieve-then-generate runs in loops, not once. That changes everything about how latency compounds and where you need to optimize.
- [Long-Running Agents: Managing Inference Costs Over Hours of Execution](https://www.generalcompute.com/blog/long-running-agents-managing-inference-costs-over-hours-of-execution.md): When an agent runs for hours instead of seconds, inference costs and context management become the dominant engineering problems. Here's how to keep them under control.
- [Evaluating Agent Performance: Latency as a First-Class Metric](https://www.generalcompute.com/blog/evaluating-agent-performance-latency-as-a-first-class-metric.md): Most agent benchmarks measure accuracy and task completion. Here's why latency belongs in your evaluation framework alongside accuracy, and how to instrument your agent to collect the data you need.
- [Browser Agents: Inference Speed as the Gating Factor](https://www.generalcompute.com/blog/browser-agents-inference-speed-as-the-gating-factor.md): Browser agents take dozens of LLM calls to complete a single web task. Here's why inference latency determines whether they're usable in practice.
- [Voice Agents and the 500ms Window: An Inference Architecture Guide](https://www.generalcompute.com/blog/voice-agents-and-the-500ms-window.md): Real-time voice conversations require end-to-end latency under 500ms. This guide breaks down the inference architecture, latency budgets, and streaming patterns needed to hit that target reliably.
- [Planning and Search with LLMs: Tree-of-Thought at Inference Scale](https://www.generalcompute.com/blog/planning-and-search-with-llms-tree-of-thought-at-inference-scale.md): Tree-of-Thought turns a single LLM call into a branching search process. Here's how it works, why inference speed determines how deep you can search, and how to build ToT agents that run fast enough to be useful.
- [The Cost of Agent Retries: When Slow Inference Kills Reliability](https://www.generalcompute.com/blog/the-cost-of-agent-retries-when-slow-inference-kills-reliability.md): Retries are how agents recover from failures, but when inference is slow, each retry multiplies latency. Here's how to think about retry budgets and why faster inference makes your agent more resilient.
- [Whisper API for Real-Time Transcription: Alternatives to OpenAI Whisper](https://www.generalcompute.com/blog/whisper-api-for-real-time-transcription-alternatives-to-openai-whisper.md): The OpenAI Whisper API works well for batch transcription but is not designed for real-time streaming. Here is how Faster-Whisper, Deepgram, and AssemblyAI compare on latency, accuracy, and pricing for live transcription use cases.
- [Fireworks AI Alternative: Faster Inference at Lower Cost](https://www.generalcompute.com/blog/fireworks-ai-alternative.md): Fireworks AI is a capable inference platform, but it has real limitations around latency, pricing at scale, and rate limits. Here's a direct comparison with alternatives and a step-by-step migration guide.
- [DeepSeek API: How to Access DeepSeek V3 Without Rate Limits](https://www.generalcompute.com/blog/deepseek-api-how-to-access-deepseek-v3-without-rate-limits.md): DeepSeek V3 is one of the strongest open-source models available, but the official API has aggressive rate limits. Here's how to access it reliably at scale, with Python and Node.js integration examples.
- [Inference API Pricing Guide: Groq vs Fireworks vs Together vs GeneralCompute](https://www.generalcompute.com/blog/inference-api-pricing-guide-groq-vs-fireworks-vs-together-vs-generalcompute.md): A practical breakdown of inference API pricing across Groq, Fireworks AI, Together AI, and GeneralCompute -- including per-token rates, hidden costs, and how to calculate your real monthly spend.
- [vLLM vs Managed Inference APIs: When to Self-Host vs Pay-as-You-Go](https://www.generalcompute.com/blog/vllm-vs-managed-inference-apis-when-to-self-host-vs-pay-as-you-go.md): A practical TCO comparison of self-hosting LLMs with vLLM versus using managed inference APIs. Includes break-even analysis and a decision framework for your team.
- [Llama API: How to Run Llama 4 via API Without Managing Infrastructure](https://www.generalcompute.com/blog/llama-api-how-to-run-llama-4-via-api.md): A practical guide to running Llama 4 through managed inference APIs. Compare providers, understand pricing and performance trade-offs, and see working integration code.
- [Best OpenAI API Alternatives in 2025: Full Developer Comparison](https://www.generalcompute.com/blog/best-openai-api-alternatives-in-2025-full-developer-comparison.md): A developer's guide to the top OpenAI API alternatives in 2025, covering GeneralCompute, Groq, Together AI, Fireworks AI, and Replicate with pricing, model selection, and migration code.
- [GeneralCompute vs Groq: Speed, Pricing, and Model Selection Compared](https://www.generalcompute.com/blog/generalcompute-vs-groq.md): A developer's comparison of GeneralCompute and Groq covering inference speed, model availability, API pricing, and when to choose each platform.
- [Scaling Laws in LLMs: What They Mean for Inference Cost in Production](https://www.generalcompute.com/blog/scaling-laws-llms-inference-cost-production.md): A practical breakdown of Chinchilla and Kaplan scaling laws, what they predict
