Providers

The API you probably already use — under Browserbase/Steel/Browserless. Listed so agents don't confuse…

Still fine for Chrome-only scripts. New agent projects should prefer Playwright unless stuck on Puppeteer.

Non-negotiable for most commercial agent products. Use Billing meters for token pass-through pricing.

Depot pick when agent artifacts get downloaded a lot. S3-compatible without the egress grief.

Depot default observability for most agent teams: open, affordable, good enough to serious.

Depot default for most agent apps under ~25–50M vectors. Don't add Pinecone until Postgres actually hurts.

Default first-pass content filter for agent I/O. Pair with Llama Guard (Groq) or NeMo Guardrails when you…

When email *must* land (receipts, auth), Postmark is the grown-up pick. Resend for DX; Postmark for…

Default object store gravity well. Consider R2/Cloudflare or DO Spaces when egress dominates.

Depot pick for serious vector workloads with filters. Self-host or Cloud — either way you can leave.

Excellent serverless Postgres for agent apps that want branching + pgvector without Supabase's full BaaS.

Depot default for app/agent transactional email. Migrate to Postmark/SendGrid only for specific…

Depot default code sandbox for agents. Use for untrusted code; Daytona when you need persistent…

Depot pick for agent-error-observability (exceptions/traces in the app hosting the agent). Keep…

Best edge runtime for lightweight agent gateways and tools. Heavy LLM loops still belong on Modal/Fly/GPU…

When your first big customer asks for Okta SSO tomorrow, WorkOS is how you say yes without a rewrite.

Depot OSS fine-tune default. Hosted LoRA → Predibase/Fireworks/Together; dead OpenPipe → do not resurrect…

Depot pick for outbound-webhooks. Use Hookdeck when the job is inbound event gateway/debugging; Courier/Knock…

Depot default for programmatic image/video/audio gen. One API to rule the media zoo.

Default STT and strong voice-agent platform. Standard Flux TTS / Voice Agent rates live since 2026-09-15…

Default serverless-edge-redis. Use classic Redis Cloud when you need always-on fat memory / Redis modules…

Default when the agent is a realtime session (voice or video) you want to own. Vapi for turnkey telephony…

Best secrets DX for most startups. Prefer Infisical when you need self-host or Agent Proxy that never hands…

Depot darling for Python agent backends and bursty GPU jobs. If you think in functions not VMs, start here.

Choose Phoenix when you want open-source observability you can own. Cloud Arize when enterprise wants a…

Depot pick when the agent needs SQL warehouse context (analytics, DuckLake/Iceberg, shared Guides). Snowflake…

Best value frontier stack and multimodal king. Use Flash for high-volume agent loops; Pro when quality…

Depot default auth for agent products with users. WorkOS when enterprise SSO is the main job.

OSS/commercial peer to LaunchDarkly/Statsig for progressive agent rollout. GrowthBook when you want…

Flags-first OSS peer. Prefer for kill-switches and audited rollouts; GrowthBook when experiments are the…

Depot default for serious coding and production agents. Prefer Sonnet for cost/quality; reserve Opus for hard…

Default OSS eval harness. Braintrust when you want hosted collaboration; Promptfoo when you want local/CI…

Depot pick for unified cloud+GPU+LLM FinOps. Helicone/Langfuse still own request-path token tracing; Vantage…

SMS/voice peer when $/part and published carrier fees matter. Twilio remains default toolkit gravity; Bird…

Inbound event-gateway peer to Svix. Prefer Hookdeck to reliably consume Stripe/GitHub/provider webhooks into…

Pick when the agent backbone must be Kafka-protocol (replay, consumers, Iceberg) without Confluent prices…

Self-host object storage when R2/Tigris/B2 SaaS is not allowed. R2 remains managed cheap-egress default.

Free edge gateway peer to Portkey/OpenRouter. Depot llm-ai-gateway pick stays Portkey for virtual…

Flags/analytics peer. Depot oss-feature-flags pick stays GrowthBook; LaunchDarkly for AgentControl; PostHog…

Depot pick for RAG rerank. Keep Cohere as general embeddings default (enterprise + multilingual maturity)…

Depot pick for a raw serverless agent event backbone (webhooks, delayed retries, fan-out) without operating…

Depot pick for agent-analytics-warehouse (event/log OLAP). MotherDuck remains sql-warehouse-agent-context for…

Depot default for agent secrets when you want OSS exit + Agent Proxy (placeholders instead of raw keys in the…

Typesense/Algolia peer for product + RAG hybrid search. Prefer Typesense when hourly HA clusters are clearer…

Top pick for production open-model serving with real DX. Prefer over random GPU DIY when you want tokens not…

Best cheap always-hot archive for datasets/checkpoints. R2 still wins agent artifact serving with unbounded…

Auth pick when agents/M2M + org SSO matter more than Clerk's consumer UI kit. Watch Twilio integration…

Best privacy-forward search API with an independent index. Strong complement to Tavily/Exa.

Default RAG metric library. Pair with Promptfoo/Braintrust for broader agent evals.

Best price/perf flags+experiments for product agents. Confirm Amplitude commercial terms at renewal…

Best when the agent is already PydanticAI/Python and you want app+LLM traces together. Langfuse stays Depot…

Depot default for self-serve GPUs. Secure Cloud for anything customer-facing; Community/spot for batch.

Depot pick for production AI gateway (fallbacks, virtual keys, budgets). OpenRouter when you want a model…

Depot default for agents that need a real browser. Pair with Stagehand/Browser Use for the reasoning loop.

Depot pick for research-style agent search. Pair with Firecrawl when you need full page content.

Depot OSS pick for AI usage metering when you refuse to marry Stripe's meter model. Still collect cash with…

Depot co-default for agent web search with Exa. Pick Tavily for classic SERP-like agent search.

Depot pick for realtime agent↔user channels (token streaming, presence, multiplayer agent UIs). Not a…

Hybrid search peer when you want Algolia-like DX without per-request fees. Sample $0.03/hr floor pinned…

Excellent for globally distributed agent sidecars and Machines-based sandboxes. Prefer Railway/Render if you…

Pick when agents need a real pub/sub fabric (multi-service, edge leafs) not just HTTP callbacks. QStash for…

Pick Steel when you need to self-host browsers or hate SaaS lock-in. Browserbase wins on polish.

Default voice layer for talking agents. Combine with Twilio for phone; watch cloning policy.

Best per-request scrape API with agent MCP. Firecrawl when you want site-wide LLM markdown; Scrapfly when you…

Enterprise product-search default. Prefer Typesense/OpenSearch when $/QPS dominates; vector DBs when pure…

Realtime TTS latency peer. Depot pick stays Cartesia; ElevenLabs for library; Hume expressive; Rime when Mist…

Default when agents need SMS/voice globally. For email-only, don't drag Twilio in.

Veteran choice for headless fleets. Great if you already speak Playwright fleets, not just agents.

Depot pick for turning the web into LLM context. Browserbase when you need to *act*, Firecrawl when you need…

Excellent Firecrawl peer when you need anti-bot + cloud browser with transparent credits. Prefer Firecrawl…

Choose when agents need sticky filesystems and full dev environments. E2B for disposable code exec. Confirm…

Depot pick for multilingual/production STT beyond English-first stacks. Deepgram remains default $/min…

Depot pick when evals are the product quality gate. Pair with Helicone/Langfuse for cheap request logs if…

Top EU list-price neocloud peer to Nebius. Prefer Nebius Depot default for global neocloud; Verda when EU…

Serverless DB peer when branching DX matters. Prefer Neon/Supabase for default Postgres+pgvector agents…

Agent-search runner when you want transparent per-call EU-leaning search. Depot pick stays Exa; Tavily…

Default when you want maximum capability and ecosystem glue. Multi-home critical paths behind OpenAI-compat…

Depot pick for document-parse (PDF/Office → LLM-ready markdown/JSON). Unstructured when you need ETL…

Best agent-native notification fabric (toolkit + human-in-the-loop). Courier remains Depot orchestration…

OSS-friendly Courier/Knock alternative. Pick when you want self-host or run-based economics; Knock for agent…

Depot pick for edge/per-tenant SQLite agent state. Serverless Postgres pick stays Neon; hot KV stays Redis…

When your vector bill scares you and much of the corpus is cold, Turbopuffer is the interesting answer.

DIY sandbox substrate with global edge. Prefer E2B unless you need custom machine control.

Depot pick for a dedicated prompt-injection firewall in front of any LLM. NeMo when you need programmable…

Depot pick when you need a scraper marketplace / reusable Actors, not just a crawl API. Firecrawl for…

GrowthBook remains Depot OSS flags pick (experimentation + flags). Flagsmith when you want flags-first…

Best when the app already ships on Vercel/AI SDK. Portkey remains production-gateway default; OpenRouter for…

Best catalog breadth for OSS models. Watch minimums and lifetime spend gates before you design around a gated…

When you need cheap always-on CPU in EU, start here. DIY the agent orchestration yourself.

Best when hybrid search and modularity matter. Cloud for convenience; self-host when lock-in anxiety spikes.

Depot pick for progressive delivery of agent prompts/models/tools in production. Statsig if you want cheaper…

OSS programmable rails peer to NeMo Guardrails. Lakera when you want a hosted injection firewall API…

Best Vapi peer for balanced inbound voice agents with clear $/min math. Vapi remains Depot default; Bland for…

Source of truth for MCP server metadata and the Depot pick for mcp-discovery since 2026-10-24…

Budget transactional pipe when you already live on AWS and own deliverability. Default agents should still…

Txn email peer when EU residency or deliverability ops matter. Depot default remains Resend; Postmark for…

Use inside GCP estates. Depot object-storage/cheap-egress defaults remain R2.

When you're truly at massive scale, Milvus/Zilliz earns its complexity. Don't start here for a chatbot MVP.

Classic simple cloud. Use for steady VMs and Spaces; pair Paperspace/RunPod for GPUs.

Convenient if you're already on OpenAI. For best image variety/cost, keep Fal in the toolkit.

R2 remains Depot object-storage / cheap-egress default (CDN+Workers gravity). Tigris when you want global…

Strong Temporal alternative for durable AI agents, especially on Workers/Vercel. Remains medium until paid…

Great TS agent product backend when you want reactive state+actions together. Prefer Supabase/Neon when…

Best when agents need behavioral messaging + txn in one vendor. Prefer Resend for pure transactional DX…

Prototype/edge RAG peer when you already on Upstash. Prefer Qdrant/Pinecone for serious filtered production…

Classic SERP API when you need engine breadth. Prefer Serper for budget Google-only; Exa/Tavily/You for…

Depot pick for research-agent enrichment pipelines. Apollo when GTM workflow glue matters more than $/record.

First-party FLUX when you want BFL terms/quality without a reseller markup. Fal remains Depot media-gen…

Depot pick for saas-oauth-integrations (embed customer SaaS OAuth/sync). Composio remains agent-tool-router…

Maintenance mode since 2026-03-03 (Mintlify); site now says 'no longer part of Mintlify'. Runner-up only, not a pick.

When your agent outage is an infra mystery, Honeycomb shines. Still pair with an LLM-native tracer for…

Nice for TS edge tools. Don't confuse isolate deploy with E2B-class untrusted code sandboxes.

GCP hyperscaler bus for agent events. Prefer Service Bus on Azure for sessions; EventBridge on AWS; QStash…

Use API for convenience or self-host SD/Flux on RunPod/Fal for control. Great escape hatch from closed image…

Best when you want connector-heavy document ETL or to self-host the OSS partitioner. LlamaParse remains the…

Promoted high after pinning official models API token matrix ($0.05/1M). Best as EU-friendly all-in-one…

Cheap hot archive peer when R2/Tigris edge story is not required. Depot cheap-egress still defaults to R2.

Serverless GPU/app peer for EU-friendly deploys. Depot neocloud pick stays Nebius; interruptible Salad…

Great when you need STT + audio intelligence features. Deepgram usually wins raw $/min STT.

Boring reliable PaaS. Great when Railway feels too hip and AWS feels too heavy.

Ship the agent UI and thin API routes here. Put long-running agent workers on Modal/Fly/Railway.

Runner-up for agent secrets when HashiCorp is already the standard. Infisical remains the depot default for…

Strong Exa peer for agent search; wins when you need Task/deep-research in one bill. Exa remains Depot…

CPaaS peer especially when Voice AI Agents + SIP streaming matter. Depot voice-agent platform pick stays…

If you live in LangGraph, LangSmith is the obvious observatory. Otherwise compare Langfuse/Helicone first.

Use when your agent/product notifies across many channels with preferences. Skip if you only send email.

Best-documented EU open-model serverless API after Mistral. Use when you want Paris residency + published €…

Depot pick for shipping a phone/voice agent without assembling vendors. LiveKit when you want OSS + realtime…

Graph memory when documents→ontology matter. Zep for temporal bipartite graphs; Mem0 for simplest hosted…

Push-first notify peer with pinned Free/$19 Growth card (Round 9 high). Prefer Knock/Courier when agents…

Depot favorite for shipping agent APIs quickly. When bills climb, move steady-state to Fly/Hetzner/DO.

Enterprise volume email incumbent. New agent apps should try Resend/Postmark first.

Use as agent hot memory/cache, not as your only 100M-vector archive.

Best DX default for durable agent workflows in TypeScript/Python. Temporal when you need harder multi-lang…

Best when agents need SVG/brand-consistent assets. Fal for general media zoo; BFL for FLUX-first.

Agent search peer with Answer/Research SKUs. Exa remains depot agent-search pick; You.com strong when ZDR +…

Strong when you need multi-turn/voice agent simulations with EU self-host. Langfuse default tracing…

Azure-estate default object store. Prefer R2/Tigris when egress-free public URLs matter.

Enterprise hybrid/vector search when already in Elastic. Prefer Typesense/Meilisearch/Algolia for product…

Default EU-friendly LLM vendor. Pair with Mistral-hosted or self-host open weights when sovereignty matters.

Depot pick for ai-merchant-of-record. Prefer Stripe when you already have tax/ops staff; OpenMeter/Kong when…

Enterprise-leaning MoR peer. Prefer Polar for AI usage products; Paddle when you want mature SaaS MoR + buyer…

Use for training/fine-tune discipline. For pure agent request tracing, Langfuse/Helicone are leaner.

Pick when you already live in Comet or want OSS Opik + optional cloud. Langfuse remains cheaper OSS default…

Azure twin to SQS/EventBridge for agent queues when already on Azure. EventBridge remains AWS hyperscaler…

Edge browser peer when already on Cloudflare. Depot cloud-browser pick stays Browserbase; Anchor/Hyperbrowser…

Budget SERP hammer. Great for hacks; prefer Tavily/Exa for cleaner long-term agent search.

Pick Orb when pricing complexity is the product (credits, tiers, simulations). Stripe meters first; Lago if…

Hyperscaler queue leaf when already on AWS. QStash remains Depot serverless HTTP bus default; SQS when…

Open-model production peer when latency/tool-calling matter. Fireworks remains depot open-model default…

Default durable orchestration for serious long-running agents. Inngest/Trigger.dev for lighter DX.

Best Scrapybara/E2B Desktop peer when you need Windows+macOS+Android fleets or local OSS sandboxes. E2B…

Use when LlamaParse Agentic still drops tables/forms. Depot still picks LlamaParse for default document-parse…

Use the free Developer tier to stand up agent evals/traces in a day; Enterprise when you need self-host…

Temporal/Inngest peer for agent task orchestration. Prefer Temporal for serious durability; Inngest for TS…

Default enterprise VM substrate. Agents should wrap it — humans shouldn't live in the console.

Strong Germany-sovereign open-model peer beside Scaleway/OVH/STACKIT. Mistral still default EU brand for…

AWS event backbone for agents (webhooks via API destinations, cron via Scheduler). Inngest/Temporal for…

EU inference bargain on published H100 €/hr. Pair with Mistral/Scaleway for model APIs; Verda when you need…

Best multi-model router for agents that need failover and experimentation. Pin providers in production; don’t…

Use for local analytics + light vector experiments. Production agent memory → pgvector/Qdrant/Mem0.

Enterprise Kafka default. Agent builders should prefer QStash (simple) or WarpStream (Kafka-compat cheaper)…

Strong usage-billing peer when Orb is too sales-opaque and you need SQL metrics + commits. OpenMeter/Kong for…

Strong custom-model serving alternative to Replicate/Modal when you bring your own model.

Best when your team already thinks in pytest/DeepEval. Braintrust remains Depot hosted-evals default…

Embed durable steps in your app's Postgres rather than running a separate workflow cluster. Temporal remains…

Enterprise proxy peer to Bright Data. Prefer Scrapfly/ZenRows/Firecrawl for agent-default scrape; Oxylabs…

Use for fetch-the-HTML jobs. For multi-step login/click agents, use Browserbase/Steel.

Use when the company already lives in 1Password Business. Don't buy 1Password just for agents — pick…

Strong enrichment peer when agents need MCP + verified contacts. PDL remains Depot pick for API-first…

Solid UK/EU GPU alternative with public card. Nebius/Verda first for neocloud; Hyperstack when UK entity +…

Best boring GPU cloud for research labs who want clusters without CoreWeave enterprise theater.

Fill the Swiss sovereign hole. Mistral remains Depot EU LLM default; Infomaniak when the RFP says Switzerland…

Enterprise Gemini path. Use AI Studio for prototypes, Vertex when security review starts asking questions.

Depot pick for enterprise A2A/MCP discovery when the fleet already lives on AWS. OpenAgora/a2a-registry.dev…

Structured extract / KG peer. Firecrawl remains llm-ready-extract default; Diffbot when you need schema'd…

Use for retrieval stacks (embed+rerank), not as your primary agent brain. Quietly excellent at enterprise RAG.

Use when latency is the product. Do not build your whole company on one Groq model SKU — they rotate.

Budget open-model workhorse and one of the cheapest large-catalog hosts. Great for evals and high-volume…

Perfect for Hub-native workflows and custom models. For hard production latency, graduate to…

Unbeatable $/quality for many workloads — if your compliance team allows it. Otherwise host similar open…

Still the zero-ops vector default. If you already run Postgres, try pgvector before paying Pinecone rents.

Best DX when you already run Stripe and need synchronous feature gates for AI usage. Metronome/Orb for…

Depot pick for mobile-computer-use on real phones. Cua if you only need Android VMs; AWS Device Farm if you…

Best open A2A discovery default for multi-vendor agents. AgentVerse if you are already Fetch.ai/uAgents…

EU multilingual STT with a now-pinned official $/hr card. Speechmatics remains Depot multilingual pick…

Depot pick for portable/user-profile memory (MCP across Claude/ChatGPT/agents). Mem0 remains general…

Budget BrowserStack/Sauce peer. Catalog as testmu-ai — do not add a second live LambdaTest SKU. Depot mobile…

Use when the agent product is a chat/feed UX. Ably remains Depot realtime channel pick for raw pub/sub token…

Solid when you're already on GCP. Otherwise Hetzner/DO/Fly are faster paths for agents.

Best when your agent product is an API gateway problem (entitlements + token meters). Orb for pricing…

Use when the team already lives in Lightning/PyTorch notebooks. Together/Unsloth remain Depot finetune…

Managed vector peer for Cassandra estates / huge write scale. Depot defaults remain Supabase pgvector /…

BrowserStack peer for real-device mobile computer-use. Prefer when you already live in Sauce QA tooling or…

Long-context runner when Gemini procurement is blocked and you want 1M ctx at $3/$15. Google AI stays Depot…

MoR peer to Polar/Paddle for digital products. Depot ai-merchant-of-record pick stays Polar for AI…

Alive. Prefer when you serve dozens of LoRA adapters on shared GPUs. Together/Fireworks remain simpler…

Monitoring/evals peer with a real $60 Premium. Prefer Braintrust/Langfuse for tracing defaults; Arthur when…

Prototype on Chroma; graduate to Qdrant/pgvector/Pinecone before your corpus and QPS grow up.

API story is no longer a mystery: self-serve keys + catalog quotes. Treat as a media aggregator (multi-vendor…

Pick for ops/RPA-replacement browser agents (invoices, portals, 2FA). Not a substitute for E2B/Cua desktops…

Managed context engine when you want memory without running a vector DB. Mem0 for simpler fact API; Cognee…

Product/lifecycle email for agent SaaS. Depot transactional default stays Resend; Postmark if deliverability…

Use when agents must render readable text in images. Fal remains default multi-model media router.

WhatsApp/SMS peer especially for EU (ex-MessageBird). Twilio remains default SMS/voice toolkit; Bird when…

Pick if Mongo is already your system of record. Otherwise pgvector/Qdrant are cleaner agent defaults.

When OpenAI Moderation taxonomy is not enough and you need programmable rails. Host Llama Guard on Groq for…

Catalog dashscope for Alibaba Model Studio / DashScope international Qwen. Complements Moonshot/MiniMax/Z.AI…

Enterprise identity Swiss Army knife. For startups, Clerk/WorkOS usually hurt less.

Pick for TS app background agent jobs. Inngest for event meshes; Temporal for polyglot durability.

Primary indie-usable Chinese multimodal API alongside Moonshot. Prefer Anthropic/OpenAI for coding agents…

Ship a model demo by lunch. Migrate hot paths to Modal/RunPod/Fireworks when the bill hurts.

Top pick for realtime low-latency voice agents. ElevenLabs for voice library breadth.

Pick when you want hosted OpenAI/Claude computer-use on stealth Chromium without owning the browser fleet…

International GLM via z.ai — catalog zai, not a second Zhipu mainland-only SKU. Strong cheap/coding runner…

Great embedded/object-store vector layer. Pair with Mem0/Zep only if you need higher-level memory APIs.

GCP twin of AWS Agent Registry. Use when agents already sit on Vertex/Gemini Enterprise; otherwise prefer…

Pick for high-volume outbound when bundled $/min and dialer features beat BYOK flexibility. Vapi/Retell when…

Solid neocloud when you want SLA + energy story. Not the first stop for hackathon GPUs.

Budget enrichment peer for agents that outgrew Clearbit-free paths. Round 9 promoted high after pinning…

Enterprise multi-model switchboard. Great when AWS is mandatory; mediocre when you just want the best Claude…

Top creative video brand. For pure agent pipelines, also price-check Fal-hosted alternatives.

Top-tier neocloud when you want list prices + serious clusters. RunPod/Vast still win for hackathon…

Use when agents need Android VMs (not real handsets). Real-device mobile CUA stays BrowserStack/Sauce/AWS…

Pick when finance wants AI unit economics and will sit through a sales cycle. Vantage for self-serve agent…

Pinned Professional $99 on the live unauthenticated card (Round 8). Strong semantic query language for agent…

Use when AWS is destiny. For greenfield agent GPUs, neoclouds win on price and speed-to-pod.

Strong EU budget open-model alternative. Mistral remains the EU product default; OVH when price + French…

Default for sales/research agents that already think in Apollo credits. Prefer People Data Labs for…

Enterprise training/inference fleet pick. Skip if you just need a single H100 for an afternoon — use RunPod.

Now legitimately self-serve for H100 bursts at published $1.99–$2.49/hr. Prefer RunPod/Nebius for indie…

Public MCP server directory; hosted URLs, API, Connect, CLI and SDKs shut down 2026-10-31.

AWS-estate default for mobile device farms. BrowserStack if you want the bigger self-serve catalog; Cua for…

Solid video gen option — A/B with Runway/Pika/Fal-hosted models for your domain.

Best when your agent already runs on Workers and you want colocated small/open models. Not a frontier brain.

Default only if the agent already runs inside AWS with IAM roles. Otherwise Doppler/Infisical.

Durable public rate card — worth A/B vs OpenAI/Anthropic on general chat. X Search now bills per…

Use when you want grounded answers without building search+cite yourself. For full agent stacks, prefer…

Interesting E2B peer for freezeable agent computers. Keep medium until preview exits; production default…

Nuclear option for hard anti-bot targets. Start with Firecrawl/ScrapingBee; escalate here when blocked.

Choose when you self-host inference on NVIDIA GPUs and want blessed containers. For serverless tokens use…

Secret weapon for cheap cloud GPUs if you tolerate OCI quirks. Not the agent-native DX winner.

MCP directory peer to Glama/Smithery — best when you need a curated sub-registry for a platform, not…

Enterprise Microsoft default. Indie agent stacks should look elsewhere first.

A/B vs Cartesia when emotional prosody is the product. Cartesia remains the low-latency default; ElevenLabs…

Pick when IT already standardized on Azure and needs VNET/private endpoints. Otherwise go direct OpenAI and…

Depot pick for tool/MCP router with managed OAuth across 1500+ apps. Official rate card is now public — Hobby…

Lightweight public/self-host A2A discovery. Prefer OpenAgora if you want LF-aligned A2A v1.0 + live testing…

Fine for experiments and small fine-tunes. Graduate to RunPod/Lambda/CoreWeave for serious training.

Fine inside GCP. Quota tickets will humble you — plan capacity early.

Default hosted agent memory bolt-on. Model retrieval volume carefully; self-host OSS if caps hurt.

Clear hosted desktop provider. E2B Desktop remains depot default if you already run E2B; Scrapybara when you…

Video-gen peer to Runway/Luma/Fal. Prefer Fal for multi-model media API DX; Kling when quality demos win and…

Viable Modal/RunPod alternative for serverless endpoints. Evaluate cold starts on your workload.

Best ultra-throughput play when your model is on Cerebras (official token matrix now pinned). Pair with…

Directory/hosting MCP peer. The Official MCP Registry is the Depot pick for mcp-discovery since…

Indie-friendly A100/H100 when you want simpler than Vast marketplace. RunPod/Salad for interruptible fleets…

Public token matrix now exists — gpt-oss-120b $0.22/$0.59 is the value SKU. Indie agents still default…

Enterprise RAG niche player. Only pick if you already standardized on AI21 or need their specific stack.

Choose when the agent needs a whole app platform, not just a code jail. E2B for pure sandboxes; Railway for…

Deploy here if your graphs already are LangGraph. Otherwise Inngest/Temporal + Langfuse to avoid stack…

Pick Maxim when pre-release agent simulation is the pain and seats are few. Braintrust for general evals…

Microsoft-shop GPU default. Everyone else should compare RunPod/Lambda first.

Useful opportunistic H100 capacity if the live console quote clears ~$2.1–2.5/hr. Remain medium until a…

Default when your agent stack already is Ray or you need serious multi-node Python. For simple serverless GPU…

Pick for production hallucination measurement + guardrails. Braintrust/Promptfoo for lighter eval loops.

Pick when agents need temporal/entity graph memory, not just vector snippets. Mem0 for simpler recall.

Budget open-model host with a pinned official matrix — strong DeepInfra/SiliconFlow peer. Prefer…

Strong self-serve neocloud with a durable public rate card and owned fleet (not a marketplace). Good…

Pick when you want an agent-native browser loop, not just a hosted Chromium. Browserbase still wins pure…

Choose when the agent IS a long-lived stateful runtime. For bolt-on memory use Mem0/Zep.

Consider in a bake-off if HoneyHive's free tier is too small and you do not want Braintrust. Do not pick…

Yes, STACKIT is real. Choose for German data-sovereignty RFPs. Not an indie default — onboarding friction is…

Metering peer when you care about per-customer AI margin, not just invoices. Vantage/CloudZero for cloud…

Enterprise EU sovereign runner when Telekom procurement wins RFPs. Not a Depot default for startups — use…

Price/catalog monster for open models. Confirm data-residency/compliance before production.

Interesting Browserbase alternative with transparent per-second rates. Trial against your session mix.

EU public-cloud GPU when Infomaniak/Mistral APIs are not enough and you must run weights in PL/DE…

Consider when Fireworks latency/price disappoints on a supported model. Keep a second host.

Best looks, worst agent API story. For programmatic image agents prefer Fal/OpenAI/Stability unless you have…

Cheap open-model inference peer to Novita/DeepInfra. Good for evals and non-critical volume.

Layer on when regulated hallucination/safety gates matter more than general CI evals. Lakera for injection…

Budget open-model overflow with a now-pinned public TEE matrix. Depot open-model production pick stays…

Catalog the current SKU (Prisma AIRS), not the dead Protect AI brand. Pick only if PANW is already the…

Price-check peer to Nebius/RunPod for on-demand H100s. Validate support and regions.

Strong when bot-detection and authenticated computer-use matter. A/B vs Browserbase/Kernel.

Pick only when your data/platform already is Databricks. Otherwise Bedrock/Vertex or direct APIs.

Use when Snowflake already holds the data. Don't drag agent backends into Snowflake just for Cortex.

Use as a router/compare layer across GPU clouds. Still pin production to a known underlying provider.

Cheap GPU peer to Vast/RunPod community. Fine for interruptible training/batch; not regulated PII.

Status check: Robust Intelligence brand is gone; buy Cisco AI Defense only inside a Cisco security standard…

Cheapest GPUs if your job can die and restart. Never put customer PII training here without eyes wide open.

Enterprise AI security peer to Prisma AIRS / Cisco AI Defense. Lakera/NeMo/Guardrails for builders…

Promoted high on official /pricing floor ($0.015/GPU-hr). Depot interruptible ultra-cheap GPU pick…

Catalog for DACH/public-sector RFPs only. Depot EU default remains Mistral; OVH/Scaleway/STACKIT for…

Only choose if your steady RPM fits inside 12–18 and flat $9–$20 beats DeepInfra/Novita per-token math. Fine…