{ "features_resolved": { "snapshot_any": true, "snapshot_mem": true, "fork": true, "pause_resume": true, "persistent_disk": true, "whole_disk": true, "persist_scope": null, "volumes": true, "auto_stop_idle": true, "long_sessions": true, "custom_image": true, "image_snapshot": true, "image_oci": true, "vm_isolation": true, "docker_inside": true, "nested_virt": true, "root": true, "systemd": true, "browser": true, "desktop": true, "computer_use": true, "code_interpreter": true, "browser_control": true, "desktop_control": true, "browser_and_desktop": true, "anti_bot": false, "captcha_solving": false, "residential_ip": false, "agent_harnesses": true, "ssh": true, "public_ipv4": false, "inbound_https": true, "custom_domain": null, "raw_tcp_inbound": null, "egress_allowlist": true, "open_internet": true, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": true, "open_source": false, "soc2": false, "hipaa": false, "sso": true, "gpu": true, "arm64": true, "windows": true, "macos": null, "lang_python": true, "lang_node": true, "lang_go": true, "lang_rust": true, "lang_java": null, "gpu_desktop": true, "credential_injection": true, "egress_http_rules": true, "wake_on_request": true, "live_resize": true, "fork_running": true, "memory_snapshot_fork": true, "webhooks": null, "scheduled_wakeups": null, "audit_logs": null, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": false, "usage_api": false, "snapshot_auto": true, "snapshot_on_demand": true, "harness_api": false, "own_agent_api": false, "fast_boot": true, "strong_isolation": true, "ingress_rules": true, "guest_firewall": true, "secret_proxy": true, "secret_proxy_any": true, "volume_attach": true, "volume_shared": true, "sidecar": null, "fuse": true, "kernel_features": true, "burst": null }, "company": { "id": "arker", "checked": "2026-09-30", "company": "Arker, Inc.", "ownership": { "value": "independent", "parent": null, "source": "https://arker.ai/about", "quote": "Investors: Backed by tier-1 venture capital firms and public company CEOs." }, "stage": { "value": "unknown", "source": "https://arker.ai/about", "quote": "Backed by tier-1 venture capital firms and public company CEOs.", "confidence": "low", "note": "Company (legal name Arker, Inc., GitHub org ArkerHQ created 2025-12-11) confirms it is VC-backed but names no specific investors, round type, amount, or date anywhere public. No TechCrunch/press coverage, no Crunchbase/PitchBook figures, and no YC page (ycombinator.com/companies/arker returns 404) or other accelerator mention were found. Too new/undisclosed to place at a specific stage without guessing." }, "accelerators": [], "total_funding_usd": { "value": null, "status": "unknown" }, "last_round": { "type": "unknown", "amount_usd": null, "date": null, "leads": [], "status": "unknown", "source": "https://arker.ai/about", "quote": "Backed by tier-1 venture capital firms and public company CEOs." }, "investors": [], "portfolio_listings": [], "revenue": { "value": null, "status": "unknown" }, "employees": { "value": 7, "source": "https://arker.ai/about" }, "note": "Very young company: GitHub org ArkerHQ created 2025-12-11; product/pricing docs only (rate card behind WorkOS login). About page names CEO Caleb Winston and 6 engineers (ex-Snowflake, AWS, NVIDIA, Stanford) and claims VC backing ('tier-1 venture capital firms and public company CEOs') but discloses no investor names, round size, or date. No TechCrunch/HN/Crunchbase coverage of funding was found; a YC company-page lookup for 'arker' returned 404. Buyers should treat funding/stage/revenue as undisclosed rather than absent." }, "id": "arker", "name": "Arker", "url": "https://arker.ai/docs/pricing", "category": "agent-sandbox", "isolation": "vm", "modes": [ { "key": "on-demand", "label": "AWS US VMs, on-demand run time (last public planning-table rates)", "flags": [], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.187736, "month_cap": null }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.372016, "month_cap": null } ], "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": null, "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "hour = 'Arker AWS US, 1 month' planning-table cost / 1,000 VMs / 730 h ($137,047 small, $271,572 large), as published on arker.ai/docs/pricing.md until at least 2026-10-01. On 2026-10-14 that table is gone: the pricing doc only says 'See arker.ai/pricing' (WorkOS login) and shows a competitor comparison. The only live public Arker number is the pools.md example quote (aws us-west-2, 32 vCPU / 64 GiB / 256 GiB, 30 days, amount_cents 146634 = $1,466.34), ~2.7% above the old-table unit rates, so these presets are kept as last-published approximations. Arker says it bills 'run time, not machine time': VMs scale to zero between runs and can be suspended by policy while waiting on outbound (e.g. LLM) responses, so use the run_time_pct knob. Only two presets were ever priced publicly; the catalog (docs/vms.md) goes to 32 vCPU / 64 GiB on CPU (ubuntu-coding-32/64; ubuntu-coding itself is 1-4 vCPU / 1-15 GiB / 32-512 GiB disk) and 28 vCPU / 256 GiB on the A100 GPU images." }, { "key": "commit-12m", "label": "AWS US pool, 12-month commitment (always-on, 19.55% off)", "flags": [ "sales" ], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.15104, "month_cap": 110.259 }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.299299, "month_cap": 218.488 } ], "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "Pools = reserved bare-metal capacity bought for a term (min 30 days), billed on the next invoice whether used or not; cannot be deleted. Private access for partnered companies. Per-instance month = the removed docs planning table's 12-month row / 1,000 (small $110.259, large $218.488; month_cap per size). Live pools are priced per request: POST /v1/pools/quote returns amount_cents, baseline_cents (on-demand) and discount_percent.", "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Discount formula D(x)=1.754933379x^3-2.505391256x^2+1.236007877x, x=(months-1)/35, from the docs pricing page as last published (gone on 2026-10-14; live discounts come from the pools quote API)." }, { "key": "commit-24m", "label": "AWS US pool, 24-month commitment (always-on, 22.83% off)", "flags": [ "sales" ], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.144871, "month_cap": 105.756 }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.287075, "month_cap": 209.565 } ], "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "As commit-12m; 24-month row. AWS has no 2-year Savings Plan, so Arker backs this with a 'procurement ladder'.", "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Planning-table 24-month row / 1,000 VMs." }, { "key": "commit-36m", "label": "AWS US pool, 36-month commitment (always-on, 48.56% off)", "flags": [ "sales" ], "os": [ "linux" ], "regions": [ "us" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.096581, "month_cap": 70.504 }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.191384, "month_cap": 139.71 } ], "pool_tiers": null, "requires_always_on": true, "always_on_month_per_instance": null, "commit_note": "As commit-12m; 36-month row.", "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "Planning-table 36-month row / 1,000 VMs." }, { "key": "eu-hetzner-proposed", "label": "Proposed Arker EU on Hetzner, 1-month (not a published offer)", "flags": [ "legacy" ], "os": [ "linux" ], "regions": [ "eu" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.030238, "month_cap": null }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.070625, "month_cap": null } ], "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": "Commit discounts 3% / 5% / 7.69% at 12/24/36 months (36-month: $20,376 / $47,591 per 1,000 VMs).", "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "NEVER PRICED (flag legacy = not on sale): the docs said \"Arker EU rows are proposed prices, not currently published offers.\" and on 2026-10-14 the rows are removed from arker.ai/docs/pricing.md entirely. Docs: 'Arker EU rows are proposed prices, not currently published offers.' $22,074 / $51,556 per 1,000 VMs per month (730 h). Priced ~30% over amortized AX42 host cost." }, { "key": "eu-scaleway-proposed", "label": "Proposed Arker EU on Scaleway, 1-month (not a published offer)", "flags": [ "legacy" ], "os": [ "linux" ], "regions": [ "eu" ], "gpu_only": false, "features": { "isolation": "vm" }, "pricing": "sizes", "vcpu_h": null, "ram_gib_h": null, "cpu_basis": "alloc", "ram_basis": "alloc", "active_floor": 0, "ram_per_vcpu": [ null, null ], "ram_fixed_per_vcpu": null, "vcpu_options": null, "min_vcpu": null, "max_vcpu": 4, "max_ram_gib": 16, "sizes": [ { "name": "small (2 vCPU / 8 GiB / 32 GiB)", "vcpu": 2, "ram_gib": 8, "disk_gib": 32, "hour": 0.030963, "month_cap": null }, { "name": "large (4 vCPU / 16 GiB / 32 GiB)", "vcpu": 4, "ram_gib": 16, "disk_gib": 32, "hour": 0.061925, "month_cap": null } ], "pool_tiers": null, "requires_always_on": false, "always_on_month_per_instance": null, "commit_note": "Commit discounts 10% / 20% / 30% at 12/24/36 months (36-month: $15,822 / $31,644 per 1,000 VMs).", "min_billed_seconds": null, "granularity_s": null, "start_fee": 0, "boot_overhead_s": null, "multiplier": 1, "windows_vcpu_h": null, "gpu": null, "note": "NEVER PRICED (flag legacy = not on sale): the docs said \"Arker EU rows are proposed prices, not currently published offers.\" and on 2026-10-14 the rows are removed from arker.ai/docs/pricing.md entirely. Proposed, not offered. $22,603 / $45,205 per 1,000 VMs per month (730 h). Markup ~50% on-demand falling to ~24% at 36 months." } ], "storage": { "included_disk_gib": 32, "disk_gib_month": null, "disk_billed_when_stopped": null, "snapshot_gib_month": null, "paused_ram_billed": null }, "network": { "egress_free_gib": null, "egress_gib": null, "ipv4_month": null, "ipv4_available": false, "internet": { "default": "configurable", "evidence": [ { "quote": "An empty rule list allows all outbound traffic and authenticated inbound access to any guest port.", "url": "https://arker.ai/docs/api/policies.md" }, { "quote": "With a non-empty rule list, unmatched outbound traffic is denied.", "url": "https://arker.ai/docs/api/policies.md" } ] } }, "plans": [ { "name": "Paid account (pay as you go)", "fee": 0, "flags": [], "fee_is_credit": false, "included_usd": 0, "per_seat": 0, "concurrency": null, "max_session_h": null, "max_vcpu": null, "max_ram_gib": null, "sizes_allowed": null, "note": "No subscription fee is mentioned anywhere public; the price catalog is behind login. No free tier ('Arker does not currently offer a free tier.', docs/deployment.md, 2026-10-14)." }, { "name": "Private region / VPC / on-prem, private networking, pools", "fee": null, "flags": [ "sales" ], "fee_is_credit": null, "included_usd": null, "per_seat": null, "concurrency": null, "max_session_h": null, "max_vcpu": null, "max_ram_gib": null, "sizes_allowed": null, "note": "Contact support@arker.ai. Pools in private access for partnered background-agent companies." } ], "free": { "monthly_credit": 0, "one_time_credit": 0 }, "features": { "linux": true, "windows": true, "macos": null, "arm64": true, "gpu": true, "gpu_types": [ "A100-80GB", "H200" ], "snapshot": "mem", "pause_resume": true, "fork": true, "persistent_disk": true, "volumes": true, "auto_stop_idle": true, "max_session_h": null, "custom_image": true, "docker_inside": true, "nested_virt": true, "root": true, "systemd": true, "browser": true, "desktop": true, "computer_use": true, "code_interpreter": true, "agent_harnesses": true, "ssh": true, "public_ipv4": false, "inbound_https": true, "custom_domain": null, "raw_tcp_inbound": null, "egress_allowlist": true, "static_egress_ip": false, "private_network": true, "self_host": false, "byoc": true, "open_source": false, "regions": [ "us", "eu" ], "soc2": false, "hipaa": false, "sso": true, "lang_python": true, "lang_node": true, "lang_go": true, "lang_rust": true, "lang_java": null, "gpu_desktop": true, "credential_injection": true, "egress_http_rules": true, "wake_on_request": true, "live_resize": true, "fork_running": true, "memory_snapshot_fork": true, "webhooks": null, "scheduled_wakeups": null, "audit_logs": null, "eu_data_residency": true, "zero_data_retention": null, "scoped_api_keys": null, "spend_limits": null, "mcp_server": false, "browser_control": true, "desktop_control": true, "anti_bot": false, "captcha_solving": false, "egress_ip_type": "datacenter", "ip_note": "VMs run in AWS regions (e.g. aws-us-west-2), so egress is cloud datacenter IPs; no proxy/residential options documented.", "isolation": "vm", "image_oci": true, "snapshot_auto": true, "snapshot_on_demand": true, "harness_api": false, "egress_allowlist_level": "both", "ingress_rules": true, "guest_firewall": true, "secret_proxy": "any-http", "volume_attach": true, "volume_shared": true, "own_agent_api": false, "max_session_unlimited": true, "fast_boot": true, "open_internet": true, "usage_api": false }, "perf": { "cpu_runs_s": null, "cold_start_ms": 75.56, "burst_tti_ms": 254.2, "dax_s": 60.53 }, "perf_source": "benchmarks.json providers.arker.computesdk: cold_start_tti median 75.56 ms (2026-08-07, 1 run), burst_tti median 254.2 ms at 100 concurrency (2026-09-25), dax 60.53 s (2026-09-25; sandbox reported 8 vCPU / 15.84 GiB aarch64, i.e. Graviton). No hpc CPU run.", "knobs": [ { "id": "run_time_pct", "label": "Share of session the VM is actually running (not scaled to zero / suspended)", "kind": "slider", "min": 0.05, "max": 1, "step": 0.05, "default": 1, "unit": "%", "options": null, "effect": "hours_factor", "modes": [ "on-demand", "eu-hetzner-proposed", "eu-scaleway-proposed" ], "why": "Arker says it charges for run time, not machine time: VMs scale to zero between runs and a scaling policy can suspend a VM while it waits on outbound (LLM) responses or between inbound requests. The exact metering rule is not public, so default is 1 (billed like wall-clock)." } ], "caveats": [ "No public list rate card: arker.ai/pricing redirects to a WorkOS login, and on 2026-10-14 the docs 'Replicas capacity planning' table (1,000 always-on VMs x 730 h) that every on-demand/commit price here is derived from is no longer on arker.ai/docs/pricing.md. The prices are the last published ones; the only live public number is the pools.md example quote ($1,466.34 for 32 vCPU / 64 GiB / 256 GiB, aws us-west-2, 30 days), ~2.7% above them.", "Only two public shapes (small 2/8/32, large 4/16/32). A 4 vCPU / 8 GiB workload is priced on the large preset ($0.372/h) because nothing public prices 4/8.", "Derived (fragile) unit rates: solving small, large and the earlier pools-doc example ($1,428 for aws us-west, 32 vCPU / 64 GiB / 256 GiB, 30 days) gave ~$0.0301/vCPU-h, ~$0.0155/GiB-h RAM, ~$0.000108/GiB-h disk (~4/8 = $0.244/h). The live example is now amount_cents 146634 ($1,466.34, docs/api/pools.md, 2026-10-14). Not used in modes.", "'Charges for run time, not machine time' is marketing text; whether idle/suspended VMs pay for memory/disk retention is not published (paused_ram_billed null). Prices exclude persistent storage; filesystem (object-storage volume) price unpublished.", "Commit modes are pools: private access, minimum 30-day term, prepaid on the next invoice, cannot be deleted; after expiry VMs keep running at normal rates. Modeled as always-on per-instance.", "EU Hetzner/Scaleway rows were explicitly \"proposed prices, not currently published offers\" and are removed from the pricing docs on 2026-10-14: flagged legacy so they are never priced.", "Egress, IPv4 and GPU prices are not published (GPU platforms x86_64-a100sxm-80gb and x86_64-h200sxm only in arker us-west per docs/regions.md, 2026-10-14; fractional vGPU from 0.125).", "No free tier. Billing granularity and minimums unknown.", "unverified: may bill run time only (active). Docs: \"charge for run time, not machine time\" (https://arker.ai/docs/pricing); the metering rule (what counts as run time, suspended-VM memory/disk) is unpublished, so modes keep alloc and the run_time_pct knob approximates it.", "Snapshot / persistent-storage price not published (\"Prices exclude persistent storage and taxes\", docs/pricing, 2026-09-28): snapshot_gib_month stays null, so retained state shows as unknown, not free.", "nested_virt is opt-in and placement-limited: fork API `nestedvirt` = \"Give an `image` or `dockerfile` VM access to `/dev/kvm`. Unsupported placements return `400`.\" (arker.ai/docs/api/fork.md, checked 2026-09-28). Which placements support it and any price difference are not published.", "Windows exists (public 'windows' base VM, 1-2 vCPU / 8 GiB, arker eu and arker us-west only, docs/vms.md, 2026-10-14) but has no published price, and the priced modes are AWS US Linux presets, so a Windows workload stays unpriced.", "macOS (`m4` platform, Tart images) is documented in the fork API but not in the public regions table or VM catalog (2026-10-14): features.macos is null (unverified), likely private access.", "Regions changed (docs/regions.md, 2026-10-14): arker eu (graviton2, zen4), arker us-west (graniterapids, A100 80GB, H200), aws eu-north-1 (new, September 2026), aws us-east-1, aws us-west-2, gcp us-central1. Arker's own eu and us-west sites publish no city or country; arker eu is mapped to DE because the DPA lists Hetzner (EEA) and AWS (United States, EEA) as hosting subprocessors, and arker us-west has no state." ], "sources": [ "https://arker.ai/docs/pricing", "https://arker.ai/docs/pricing.md", "https://arker.ai/pricing", "https://arker.ai/docs/api/pools.md", "https://arker.ai/docs/api/policies.md", "https://arker.ai/docs/deployment.md", "https://arker.ai/docs/regions.md", "https://arker.ai/docs/changelog.md", "https://arker.ai/docs/api/fork.md", "https://arker.ai/docs/desktops.md", "https://arker.ai/docs/vms.md", "https://arker.ai/dpa" ], "feature_evidence": { "snapshot_auto": { "value": true, "source": "https://arker.ai/docs/api/filesystems.md", "quote": "Every VM's memory and disk are persistent by default on local NVMe and on remote object storage when idle.", "note": "Trigger: VM goes idle (no active run) or a scaling policy suspends it (e.g. while waiting on an outbound LLM response); resumes automatically on next run / inbound request. Memory (running processes) included. Intro also says Arker can 'transparently snapshot machines at high frequency and restore to other hosts in the event of failure' (https://arker.ai/docs.md) and 'durable' VMs make every run a transaction.", "checked": "2026-09-29" }, "snapshot_on_demand": { "value": true, "source": "https://arker.ai/docs/api/fork.md", "quote": "Omit it or pass `[\"disk\", \"memory\"]` for a warm fork that resumes the source filesystem and running processes.", "note": "No separate named-snapshot endpoint: the checkpoint primitive is POST /v1/fork of any of your VMs (source_vm_id), which copies disk + memory into a new VM that can itself be forked again (template), or disk-only with layers ['disk']. Memory included; new machines boot from it. Can be marked public for other orgs. No explicit 'restore in place' to an earlier point.", "checked": "2026-09-29" }, "harness_api": { "value": false, "source": "https://arker.ai/docs/api/overview.md", "quote": "The Arker API is a small set of operations over machines and filesystems.", "note": "Full operation list = fork, vms, policies, runs (generic command exec with durable runs, sessions, PTY), sync, mounts, filesystems. Agents (Claude Code, Codex, OpenCode, Cursor Agent, Aider) are only preinstalled in the ubuntu-coding public VM ('sign in with your own credentials'), no agent-level prompt/stream/resume API.", "checked": "2026-09-29" }, "egress_allowlist": { "value": true, "source": "https://arker.ai/docs/api/policies.md", "quote": "With a non-empty rule list, unmatched outbound traffic is denied.", "note": "Per-VM policy document (PUT /v1/vms/{id}/policies), first-match rules on hosts (SNI/Host suffix match), IPs/CIDRs and ports; 'Layer-4 rules (ips/ports/hosts) are enforced in kernel' on the host side. Deny-all egress with a single rule. Added July 2026.", "checked": "2026-09-29", "level": "both" }, "egress_http_rules": { "value": true, "source": "https://arker.ai/docs/api/policies.md", "quote": "layer-7 criteria and the `rewrite`/`gate`/`scaling` actions run through a managed proxy.", "note": "L7 match on methods, paths (segment prefix), headers, body_contains; actions allow/deny/rewrite (host, path, headers, body)/gate (external HTTP authorization call per request, fail-closed by default).", "checked": "2026-09-29" }, "ingress_rules": { "value": true, "source": "https://arker.ai/docs/api/policies.md", "quote": "Only ports in an explicit inbound allow rule are exposed.", "note": "Inbound rules pick exposed ports, auth 'arker' (Bearer API key of the owning org, default) or 'open' (public); match on caller source IPs/CIDRs. URL form -.-.arker.app.", "checked": "2026-09-29" }, "guest_firewall": { "value": true, "source": null, "quote": null, "note": "Full VM (custom hypervisor) with root and systemd; not stated in docs.", "checked": "2026-09-29", "derived": true }, "secret_proxy": { "value": "any-http", "source": "https://arker.ai/docs/api/policies.md", "quote": "Keep an Anthropic API key out of the machine: match calls to `api.anthropic.com` and set the key from a rewrite. Code in the machine calls Anthropic with no credentials; the proxy injects it.", "note": "Outbound rewrite rules work on any host match and can set headers, path, host or body with ${secret:NAME} named secrets (encrypted at rest, redacted on read). HTTP(S) only via the managed L7 proxy; no mention of databases/SSH/git-over-SSH.", "checked": "2026-09-29" }, "volume_attach": { "value": true, "source": "https://arker.ai/docs/concepts.md", "quote": "A filesystem is a standalone, persistent, network-attached, object-storage-backed volume that outlives any single machine.", "note": "POST /v1/filesystems then POST /v1/vms/{id}/mounts at a path. Price unpublished.", "checked": "2026-09-29" }, "volume_shared": { "value": true, "source": "https://arker.ai/docs/api/filesystems.md", "quote": "so one filesystem can be mounted into many machines. Reads and writes under the path land in the shared volume.", "note": "read-write, object-storage-backed network filesystem; consistency semantics not documented.", "checked": "2026-09-29" }, "docker_inside": { "value": true, "source": "https://arker.ai/docs/vms.md", "quote": "Devtools Docker CLI, Git and Git LFS, GitHub CLI, SSH", "note": "Docker CLI is preinstalled on Arker's coding/dev VM images. Machines are full VMs with root, so a container runtime works normally (no nested virtualization required for containers, unlike VMs-in-VM).", "checked": "2026-09-30" }, "captcha_solving": { "value": false, "source": "https://arker.ai/docs/browsers.md", "quote": "There are at least two valid reasons to run browsers on hyper-elastic VMs instead of Browserbase/Kernel/similar: One, the freedom to fork Chromium and optimize... for stealth", "note": "not offered: outside the product's scope. Arker sells general-purpose VMs/hypervisor primitives (fork/run/sync) with a browser-capable base image; it documents anti-bot-style Chromium tuning as something you build yourself, not a built-in CAPTCHA-solving service. No CAPTCHA feature is mentioned anywhere in the docs.", "checked": "2026-09-30" }, "fork_running": { "value": true, "source": "https://arker.ai/docs/api/fork.md", "quote": "State to inherit from a source VM. Omit it or pass [\"disk\", \"memory\"] for a warm fork that resumes the source filesystem and running processes.", "note": "The fork API takes a memory+disk snapshot of the source VM (state can be 'running' per the VM object) and boots a new machine with its processes resumed; nothing in /docs/api/fork or /docs/api/vms requires stopping/idling the source first, and 'hyper-elastic, durable VMs' with fork as a first-class primitive is marketed as a live-branching feature.", "checked": "2026-09-30" }, "live_resize": { "value": true, "source": "https://arker.ai/docs/api/vms.md", "quote": "Update a machine's resources, SSH keys, or network policy. `resources` Resource changes: { \"vcpu\": 4, \"memory_mib\": 8192, \"disk_mib\": 16384 }.", "note": "PATCH /v1/vms/{id} changes vCPU/RAM/disk on an existing machine in place; POST /v1/vms/{id}/runs also accepts vcpu_count/memory_mib/disk_mib to 'resize the VM before running the command', values persisting after the run.", "checked": "2026-09-30" }, "own_agent_api": { "value": false, "source": "https://arker.ai/docs/api/overview.md", "quote": "The Arker API is a small set of operations over machines and filesystems.", "note": "not offered: outside the product's scope. Arker's API covers fork, vms, run, sync, filesystems, policies, pools — infrastructure primitives — with no first-party browser/computer-use/research agent that Arker itself operates and exposes via API.", "checked": "2026-09-30" }, "mcp_server": { "value": false, "source": "https://arker.ai/docs/api/overview.md", "quote": "Arker exposes a REST API, a CLI, a web UI, and Python and TypeScript clients.", "note": "No MCP server or MCP mention anywhere in the docs, llms.txt, changelog, or site search; the listed integration surfaces are REST, CLI, web console, and Python/TypeScript SDKs only.", "checked": "2026-09-30" }, "public_ipv4": { "value": false, "source": "https://arker.ai/docs/concepts.md", "quote": "A machine is reached through the `.app` data plane. Guest port `8080` uses the base hostname `.-.arker.app`; another port uses `-.-.arker.app`.", "note": "Inbound reachability is via Arker's managed relay/hostname system (arker.app), gated by policy auth ('arker' bearer-token or 'open'), not a per-machine public IPv4 address. Matches card's network.ipv4_available: false.", "checked": "2026-09-30" }, "static_egress_ip": { "value": false, "source": "https://arker.ai/docs/regions.md", "quote": "| `aws` | `us-west-2` | `https://aws-us-west-2.arker.ai/` | `graviton2`, `graviton4` |", "note": "VMs run in underlying AWS/GCP/Arker-bare-metal regions with ordinary datacenter egress (card's egress_ip_type: datacenter); no static/dedicated outbound IP option is documented anywhere in the API, policies, or pricing docs.", "checked": "2026-10-14" }, "soc2": { "value": false, "source": "https://arker.ai/dpa", "quote": "Arker's most recent SOC 2 report or, until one is issued, a summary of Arker's security controls and the status of its SOC 2 audit", "note": "The DPA's own wording ('until one is issued') indicates Arker has not yet completed a SOC 2 attestation as of 2026-09-25; it commits to sharing audit status/security-control summaries instead.", "checked": "2026-09-30" }, "hipaa": { "value": false, "source": "https://arker.ai/dpa", "quote": "Customer will not submit payment card data subject to [PCI DSS], or protected health information subject to the U.S. Health Insurance Portability and Accountability Act, unless the parties have signed a separate written agreement covering that data.", "note": "No standing BAA / HIPAA-eligibility program is advertised; PHI is disallowed by default and only permitted under a separate bespoke agreement, so there is no generally available HIPAA offering.", "checked": "2026-09-30" }, "sso": { "value": true, "source": "https://arker.ai/terms", "quote": "Accounts may be authenticated using Arker's credentials or any third-party authentication or single sign-on provider that Arker may support from time to time.", "note": "The DPA's subprocessor list also names WorkOS with function 'Authentication and single sign-on'.", "checked": "2026-09-30" }, "max_session_h": { "value": "none", "source": "https://arker.ai/docs/api/vms.md", "quote": "state: \"idle\" | \"running\" — Lifecycle state.", "note": "Arker VMs are persistent machines (fork once, then run/patch/delete at will) with no documented maximum session/run duration; auto_stop_idle governs idling, not a hard session-length cap, and no timeout is mentioned in vms.md, run.md, or pricing docs. Treated like a VM product with no imposed ceiling.", "checked": "2026-09-30" }, "windows": { "value": true, "source": "https://arker.ai/docs/vms.md", "quote": "| `windows` | 1–2 | 8 | 8–512 | | `eu`, `us-west` | `graniterapids`, `zen4` |", "note": "Public ArkerHQ base VM 'windows' (1-2 vCPU, 8 GiB, 8-512 GiB disk) in the arker eu and arker us-west regions only, not on the AWS/GCP placements the priced modes use. Policies doc confirms Windows guests: 'HTTP rules work only on ports `80` and `443`, and not on Windows.' Windows price not published.", "checked": "2026-10-14" }, "macos": { "value": null, "source": "https://arker.ai/docs/api/fork.md", "quote": "On `m4` (macOS), pass a Tart macOS image; it boots at 4 vCPUs and 8192 MiB unless `resources` says otherwise.", "note": "The fork API documents an `m4` macOS platform (Tart images; 'only Arker's base VMs ... fork warm'), but no `m4` platform is in the public regions table (docs/regions.md) and no macOS VM is in the public catalog (docs/vms.md). Regions page: 'For access to compute on a particular provider or region, contact us.' So it is unreleased or private access; unverified, not offered self-serve.", "checked": "2026-10-14" }, "self_host": { "value": false, "source": "https://arker.ai/docs/deployment.md", "quote": "Private regions may be created for a VPC or on-prem hardware cluster. The control plane runs in the cloud and ensures hyper-elastic and durable orchestration while the private region provides raw compute.", "note": "Even the 'private region' / on-prem option keeps the control plane on Arker's own cloud (contact-sales only); customers cannot run the whole Arker service themselves. No self-hostable Arker build is published; deployment.md only points to BLAST, a separate open-source project.", "checked": "2026-10-14" }, "open_source": { "value": false, "source": "https://arker.ai/docs/deployment.md", "quote": "BLAST — open-source VMs as a service", "note": "Under 'Free tier' the docs only link BLAST (github.com/stanford-mast/blast), a separate open-source project, then say 'Arker does not currently offer a free tier.' No Arker source code or open-source edition is published; the older 'full open-source version of Arker' wording was removed.", "checked": "2026-10-14" }, "fast_boot": { "value": true, "source": "https://www.computesdk.com/benchmarks", "quote": "burst of 100: p95 267 ms, 100% started", "note": "ComputeSDK sandbox benchmark 2026-09-25", "checked": "2026-09-25" }, "usage_api": { "value": false, "source": "https://arker.ai/docs/api/overview.md", "quote": "The Arker API is a small set of operations over machines and filesystems. Operations: fork, vms, policies, runs, sync, mounts, filesystems. No billing or usage endpoint is documented.", "note": "granularity: account-only; cost: no; The API provides fork, vms, policies, runs, sync, and filesystem operations. Run/VM objects return timing (started_at, completed_at, created_at, last_active_at) but no billable usage data. The pricing page is behind a login (arker.ai/console/pricing). Arker charges for run time with scaling policies, but the exact metering rule is unpublished, so billable usage cannot be calculated from API data alone.", "confidence": "high", "checked": "2026-10-01" }, "gpu_types": { "value": [ "A100-80GB", "H200" ], "source": "https://arker.ai/docs/regions.md", "quote": "| `arker` | `us-west` | `https://arker-us-west.arker.ai/` | `graniterapids`, `x86_64-a100sxm-80gb`, `x86_64-h200sxm` |", "note": "Only GPU platforms in the public regions table; A100-40GB and H100 are no longer listed. Public GPU images (ubuntu-gpu, ubuntu-gpu-small) run on x86_64-a100sxm-80gb only. Fork API: `vgpu` accepts increments of 0.125 from 0.125 to 1. GPU price not published.", "checked": "2026-10-14" } }, "locations": [ { "country": "US", "state": "OR", "city": "Portland", "provider_region": "aws-us-west-2", "self_serve": true, "choosable": true }, { "country": "US", "state": "VA", "city": null, "provider_region": "aws-us-east-1", "self_serve": true, "choosable": true }, { "country": "SE", "state": null, "city": "Stockholm", "provider_region": "aws-eu-north-1", "self_serve": true, "choosable": true }, { "country": "US", "state": "IA", "city": "Council Bluffs", "provider_region": "gcp-us-central1", "self_serve": true, "choosable": true }, { "country": "DE", "state": null, "city": null, "provider_region": "arker-eu", "self_serve": true, "choosable": true }, { "country": "US", "state": null, "city": null, "provider_region": "arker-us-west", "self_serve": true, "choosable": true } ], "locations_source": "https://arker.ai/docs/regions.md", "locations_quote": "| `arker` | `eu` | `https://arker-eu.arker.ai/` | `graviton2`, `zen4` | ... | `aws` | `eu-north-1` | `https://aws-eu-north-1.arker.ai/` | `graviton2` | | `aws` | `us-east-1` | `https://aws-us-east-1.arker.ai/` | `graviton4`, `icelake` | | `aws` | `us-west-2` | `https://aws-us-west-2.arker.ai/` | `graviton2`, `graviton4` | | `gcp` | `us-central1` | `https://gcp-us-central1.arker.ai/` | `sapphirerapids` |", "locations_checked": "2026-10-14" }