What changed, and when

Follow newly published and updated guides, site releases, model and pricing updates, and edits to tracked provider values. Model updates appear when approved data is deployed; pending imports are not published changes.

November 10, 2026 Catalog last touched
32 Products tracked
156 Recorded edits
Live database Serving from

New and updated guides

Published guide releases, newest first. Drafts and scheduled guides appear only after publication.

Site releases and model updates

Other publication items from the latest 300 recorded updates. Model comparisons start with the September 16, 2026 release. Counts describe changes to published listings, not necessarily vendor launch dates.

Provider field edit history

  1. 6 changes

  2. 32 changes

    • Agent Router GitHub stars automated refresh
      2112 2341

      Verified GitHub repository ID 875880554 after following its explicit same-origin redirect. Same repository, now theagentrouter/agent-router; 2,341 stars, Apache-2.0, not archived. Checked 2026-11-03.

      Source for this change ↗
    • Envoy AI Gateway name automated refresh
      Envoy AI Gateway Agent Router

      Official name change from Envoy AI Gateway. Same code and maintainers, not a new provider; retain the envoy-ai-gateway catalog slug and links.

      Source for this change ↗
    • Envoy AI Gateway product fate automated refresh
      Actively developed and freshly stable. v1.0.0 shipped 23 June 2026 as the first GA release with a committed-stable `v1beta1` control-plane API; v1.1.0 followed on 21 August 2026, and the repository was last pushed 2 September 2026 with 286 open issues and 1,987 stars ([v1.0 release notes](https://aigateway.envoyproxy.io/release-notes/v1.0); [releases API](https://api.github.com/repos/envoyproxy/ai-gateway/releases/latest); [GitHub API](https://api.github.com/repos/envoyproxy/ai-gateway), 2026-09-02). Cadence tracks Envoy Gateway - "approximately every 2-3 months" - with end-of-life two releases after a version ships, and the community meets weekly on the Envoy Slack ([RELEASES.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/RELEASES.md); [README](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/README.md)). Governance is multi-vendor: maintainers from Tetrate, Bloomberg, Tencent, Netflix and Nutanix, several of whom also maintain Envoy Proxy, Envoy Gateway, KServe or sit on Kubeflow's steering committee ([MAINTAINERS.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/MAINTAINERS.md); [1.0 announcement](https://aigateway.envoyproxy.io/blog/v1.0-release-announcement)). Key risk to watch: `QuotaPolicy` is still `v1alpha1`-only and therefore outside the stability guarantee, with its `serviceQuota` field "accepted but not yet enforced end-to-end" ([RELEASES.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/RELEASES.md)). Formerly Envoy AI Gateway, now Agent Router, an Agentic AI Foundation project. The official announcement confirms the same code, maintainers and Apache 2.0 license. The CLI remains aigw, the CRDs and aigateway.envoyproxy.io API group are unchanged, and existing manifests require no rename migration. The repository moved to theagentrouter/agent-router and the website/docs to theagentrouter.ai. GitHub repository ID 875880554 is not archived. Verified 2026-11-03. Historical research from 2026-09-02 (not revalidated during this manual refresh): Actively developed and freshly stable. v1.0.0 shipped 23 June 2026 as the first GA release with a committed-stable `v1beta1` control-plane API; v1.1.0 followed on 21 August 2026, and the repository was last pushed 2 September 2026 with 286 open issues and 1,987 stars ([v1.0 release notes](https://aigateway.envoyproxy.io/release-notes/v1.0); [releases API](https://api.github.com/repos/envoyproxy/ai-gateway/releases/latest); [GitHub API](https://api.github.com/repos/envoyproxy/ai-gateway), 2026-09-02). Cadence tracks Envoy Gateway - "approximately every 2-3 months" - with end-of-life two releases after a version ships, and the community meets weekly on the Envoy Slack ([RELEASES.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/RELEASES.md); [README](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/README.md)). Governance is multi-vendor: maintainers from Tetrate, Bloomberg, Tencent, Netflix and Nutanix, several of whom also maintain Envoy Proxy, Envoy Gateway, KServe or sit on Kubeflow's steering committee ([MAINTAINERS.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/MAINTAINERS.md); [1.0 announcement](https://aigateway.envoyproxy.io/blog/v1.0-release-announcement)). Key risk to watch: `QuotaPolicy` is still `v1alpha1`-only and therefore outside the stability guarantee, with its `serviceQuota` field "accepted but not yet enforced end-to-end" ([RELEASES.md](https://raw.githubusercontent.com/envoyproxy/ai-gateway/main/RELEASES.md)).

      Prepend the verified rebrand/current project identity, while retaining prior sourced operating caveats as explicitly historical research. No prior limitation is assumed resolved.

      Source for this change ↗
    • Envoy AI Gateway url docs automated refresh
      https://aigateway.envoyproxy.io/docs/ https://theagentrouter.ai/docs/

      New official documentation fetched successfully (HTTP 200), explicitly describing Agent Router as formerly Envoy AI Gateway.

      Source for this change ↗
    • Envoy AI Gateway url home automated refresh
      https://aigateway.envoyproxy.io/ https://theagentrouter.ai/

      Official project announcement names the new canonical website and states the old domain redirects.

      Source for this change ↗
    • Envoy AI Gateway url github automated refresh
      https://github.com/envoyproxy/ai-gateway https://github.com/theagentrouter/agent-router

      Official rebrand announcement confirms this repository move; the old GitHub API URL returns HTTP 301, causing the automated check failure. Existing catalog slug is preserved.

      Source for this change ↗
    • Respan Models available automated refresh
      741 726

      Manual recheck after HTTP 429: a fresh successful fetch of the complete official models array returned 726 active entries with 726 distinct IDs and no API error. This is an endpoint-entry count, not unique base models or account entitlement. Checked 2026-11-03.

      Source for this change ↗
    • Braintrust Gateway url docs automated refresh
      https://www.braintrust.dev/docs/deploy/gateway https://docs.braintrust.dev/deploy/gateway

      Canonical current Gateway documentation after www.braintrust.dev redirects to docs.braintrust.dev.

      Source for this change ↗
    • Portkey model count high automated refresh
      2300 Not published

      The old 2,300 upper bound conflates a separate pricing registry with gateway model routing and no source establishes a ceiling.

      Source for this change ↗
    • Orq.ai Router url docs automated refresh
      https://docs.orq.ai/docs/ai-gateway https://docs.orq.ai/ai-gateway/get-started/introduction

      Canonical gateway introduction after the old /docs/ai-gateway path redirects; verified HTTP 200.

      Source for this change ↗
    • Orq.ai Router model count note automated refresh
      Vendor-stated 500+ models from 30+ providers on the pricing page; the docs cite 300+ models and a competitor comparison cites 400+ from 28+ providers, so the number moves by page. Current pricing and gateway introduction state 500+ models; this is a vendor-published lower bound, not an exact 500-model count or an exhaustive account inventory. Exact availability depends on the configured workspace and providers. Reviewed 2026-11-03.

      Current pricing and gateway introduction state 500+ models; this is a vendor-published lower bound, not an exact 500-model count or an exhaustive account inventory. Exact availability depends on the configured workspace and providers. Reviewed 2026-11-03.

      Source for this change ↗
    • Orq.ai Router model count high automated refresh
      500 Not published

      The current 500+ headline supplies no finite upper bound. Keep the existing lower bound; do not show 500 as an exact total.

      Source for this change ↗
    • Helicone access meta automated refresh
      {"qualified":[],"model_count_kind":"vendor_total","provider_count_kind":"vendor_total"} {"qualified":[],"model_count_kind":"enumerated","provider_count_kind":"vendor_total"}

      The count comes from the successfully retrieved public registry rather than a vendor marketing headline. Preserve unrelated access metadata.

      Source for this change ↗
    • Helicone model count source note automated refresh
      "1 API. 100+ models" ([GitHub](https://github.com/Helicone/ai-gateway)); "100+ Models Available" ([Integrations overview](https://docs.helicone.ai/gateway/integrations/overview)). Both undated. The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      Source for this change ↗
    • Helicone model count high automated refresh
      100 111

      The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      Source for this change ↗
    • Portkey GitHub stars automated refresh
      13136 13315

      Manual GitHub API verification on 2026-11-03: 13315 stars; license unchanged (MIT); archived=false.

      Source for this change ↗
    • Kong AI Gateway GitHub stars automated refresh
      44100 44347

      Manual GitHub API verification on 2026-11-03: 44347 stars; license unchanged (Apache-2.0); archived=false.

      Source for this change ↗
    • Helicone GitHub stars automated refresh
      6241 6259

      Manual GitHub API verification on 2026-11-03: 6259 stars; license unchanged (Apache-2.0); archived=false.

      Source for this change ↗
    • Braintrust Gateway GitHub stars automated refresh
      409 411

      Manual GitHub API verification on 2026-11-03: 411 stars; license unchanged (MIT); archived=false. This linked repository is the legacy Braintrust AI proxy, not proof of the current hosted Gateway implementation or license.

      Source for this change ↗
    • Portkey model count note automated refresh
      Docs and the Gateway 2.0 announcement state a unified interface to 250+ models; the gateway's own /v1/models endpoint requires an API key, so no public catalog endpoint exists. The gateway README still contains 250+ and 1,600+ routing claims, while its separate 2,300+ claim describes the open-source pricing registry. Retain a conservative 250+ headline floor with no finite upper bound. These claims are not an authenticated routing inventory. Reviewed 2026-11-03.

      The gateway README still contains 250+ and 1,600+ routing claims, while its separate 2,300+ claim describes the open-source pricing registry. Retain a conservative 250+ headline floor with no finite upper bound. These claims are not an authenticated routing inventory. Reviewed 2026-11-03.

      Source for this change ↗
    • Portkey model count source note automated refresh
      Vendor numbers disagree: "over 250 AI models" ([Portkey docs](https://portkey.ai/docs/introduction/what-is-portkey), undated), "1,600+ LLMs" ([Portkey hybrid deployment](https://portkey.ai/docs/enterprise/hybrid2), 2024-10-01), "2,300+ models" ([portkey-ai/gateway](https://github.com/portkey-ai/gateway), undated). The gateway README still contains 250+ and 1,600+ routing claims, while its separate 2,300+ claim describes the open-source pricing registry. Retain a conservative 250+ headline floor with no finite upper bound. These claims are not an authenticated routing inventory. Reviewed 2026-11-03.

      The gateway README still contains 250+ and 1,600+ routing claims, while its separate 2,300+ claim describes the open-source pricing registry. Retain a conservative 250+ headline floor with no finite upper bound. These claims are not an authenticated routing inventory. Reviewed 2026-11-03.

      Source for this change ↗
    • Orq.ai Router model count source note automated refresh
      "500+ models" ([Supported models](https://docs.orq.ai/docs/ai-gateway/supported-models)), page undated. Current pricing and gateway introduction state 500+ models; this is a vendor-published lower bound, not an exact 500-model count or an exhaustive account inventory. Exact availability depends on the configured workspace and providers. Reviewed 2026-11-03.

      Current pricing and gateway introduction state 500+ models; this is a vendor-published lower bound, not an exact 500-model count or an exhaustive account inventory. Exact availability depends on the configured workspace and providers. Reviewed 2026-11-03.

      Source for this change ↗
    • Helicone model count note automated refresh
      Live public model registry returned 111 models spanning 21 provider endpoints on 2026-08-29; docs describe the gateway as 100+ models / 100+ providers. The simpler https://ai-gateway.helicone.ai/v1/models endpoint returns the same 111 models without pricing. The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      Source for this change ↗
    • Helicone model count low automated refresh
      100 111

      The public model registry returned 111 top-level model records with 111 distinct IDs, data.total=111 and error=null on 2026-11-03. Count is registry model entries, not provider-route variants, unique base models or proof of account availability. Checked manually; automated model checks remain disabled.

      Source for this change ↗
    • Requesty Models available automated refresh
      802 808

      Manual fresh public models API enumeration on 2026-11-03; all returned IDs are distinct. Count is endpoint entries including aliases/variants, not unique base models or account entitlement. OpenRouter total_count matches data length and links.next is null.

      Source for this change ↗
    • OpenRouter Models available automated refresh
      458 454

      Manual fresh public models API enumeration on 2026-11-03; all returned IDs are distinct. Count is endpoint entries including aliases/variants, not unique base models or account entitlement. OpenRouter total_count matches data length and links.next is null.

      Source for this change ↗
    • Braintrust Gateway product fate automated refresh
      Actively developed, with an $80M Series B led by ICONIQ announced in February 2026 and new dashboard features in August 2026. Manual scope check on 2026-11-03: Braintrust now marks the legacy AI proxy deprecated and not regularly maintained. Current hosted Gateway docs describe public preview, free use and no hosted Gateway SLA. Existing Braintrust-hosted /v1/proxy requests route to the new Gateway. The linked braintrust-proxy GitHub repository, its MIT license and its historical 100+ model claim describe the legacy proxy; a complete current-Gateway account inventory was not verified. This does not mean the Braintrust product or current Gateway is abandoned. Historical company research (not revalidated in this manual refresh): Actively developed, with an $80M Series B led by ICONIQ announced in February 2026 and new dashboard features in August 2026.

      Distinguish deprecation of the legacy proxy from the current Gateway. Preserve previous company research as explicitly historical rather than implying the entire product is deprecated.

      Source for this change ↗
    • Hugging Face Inference Providers Models available automated refresh
      136 123

      Live endpoint returned 123 models (was 136, a change of -13). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Eden AI Models available automated refresh
      1207 1291

      Live endpoint returned 1291 models (was 1207, a change of +84). Counted directly from the vendor's public catalog API.

      Source for this change ↗
    • Vercel AI Gateway Models available automated refresh
      482 497

      Live endpoint returned 497 models (was 482, a change of +15). Counted directly from the vendor's public catalog API.

      Source for this change ↗
    • agentgateway GitHub stars automated refresh
      5680 5922

      GitHub reports 5,922 stars (was 5,680).

      Source for this change ↗
    • Bifrost GitHub stars automated refresh
      8972 9331

      GitHub reports 9,331 stars (was 8,972).

      Source for this change ↗
  3. 5 changes

  4. 10 changes

  5. 1 change

  6. 11 changes

  7. 9 changes

    • Velokey Models available automated refresh
      120 94

      Live endpoint returned 94 models (was 120, a change of -26). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Merge Gateway Models available automated refresh
      274 3

      Live endpoint returned 3 models (was 274, a change of -271). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Higress GitHub stars automated refresh
      9273 9458

      GitHub reports 9,458 stars (was 9,273).

      Source for this change ↗
    • Together AI Models available automated refresh
      200 272

      Live endpoint returned 272 models (was 200, a change of +72). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Groq Models available automated refresh
      Not published 11

      Live endpoint returned 11 models (was 0, a change of +11). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Fireworks AI Models available automated refresh
      100 27

      Live endpoint returned 27 models (was 100, a change of -73). Counted directly from the vendor's authenticated account catalog API.

      Source for this change ↗
    • Eden AI Models available automated refresh
      1066 1104

      Live endpoint returned 1104 models (was 1066, a change of +38). Counted directly from the vendor's own models API.

      Source for this change ↗
    • OpenRouter Models available automated refresh
      445 458

      Live endpoint returned 458 models (was 445, a change of +13). Counted directly from the vendor's own models API.

      Source for this change ↗
    • Bifrost GitHub stars automated refresh
      8142 8359

      GitHub reports 8,359 stars (was 8,142).

      Source for this change ↗
  8. 5 changes

  9. 1 change

    • Velokey provider added manual edit
      Not published Managed marketplace

      Owner-requested onboarding; retention conflict and unknown fees preserved. Static publication requires deployment.

      Source for this change ↗
  10. 12 changes

  11. 7 changes

    • Vercel AI Gateway Distributed tracing automated refresh
      not_documented otel

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Vercel AI Gateway tracing note automated refresh
      `n.a.` — No OpenTelemetry or agent tracing documented for the gateway ([Observability](https://vercel.com/docs/ai-gateway/observability-and-spend/observability)) Native OTLP/HTTP Trace Drains on Pro and Enterprise; separate delivery and egress charges. Request, routing, model-attempt and provider spans contain metadata, not prompt or completion content. Request traces do not establish complete application or agent tracing.

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Vercel AI Gateway Seat fee automated refresh
      20 0

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Vercel AI Gateway Pricing model detail automated refresh
      Credit-based pay-as-you-go with 0% markup and no platform fee on tokens. Monetizes gateway features à la carte: provider allowlists, ZDR, custom reporting writes/queries, Trace Drains. Enterprise can pay by invoice with no processing fees. Pricing page last updated 2026-08-23. Gateway credits are pay-as-you-go with no token markup or mandatory platform subscription. Optional governance, reporting and trace exports have separate charges; some require Pro or Enterprise. Include only incremental platform costs for the selected features.

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Vercel AI Gateway extra fees automated refresh
      [{"label":"Team-wide zero data retention","amount":"$0.10 per 1,000 requests (Pro and Enterprise)"},{"label":"Team-wide provider allowlist","amount":"$0.10 per 1,000 successful requests (Pro and Enterprise)"},{"label":"Custom Reporting writes","amount":"$0.075 per 1,000 tag / user ID / quota entity writes"},{"label":"Custom Reporting queries","amount":"$5 per 1,000 queries to the reporting endpoint"},{"label":"Vercel Pro developer seat","amount":"$20 per month"}] [{"label":"Team-wide zero data retention","amount":"$0.10 per 1,000 requests (Pro and Enterprise)"},{"label":"Team-wide provider allowlist","amount":"$0.10 per 1,000 successful requests (Pro and Enterprise)"},{"label":"Custom Reporting writes","amount":"$0.075 per 1,000 tag / user ID / quota entity writes"},{"label":"Custom Reporting queries","amount":"$5 per 1,000 queries to the reporting endpoint"},{"label":"Optional Vercel Pro developer seat","amount":"$20 per month"},{"label":"Optional Trace Drains delivery","amount":"$0.05 per 1,000 traces (Pro and Enterprise)"},{"label":"Optional Trace Drains egress","amount":"$0.50 per GB (Pro and Enterprise)"}]

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Cloudflare AI Gateway Distributed tracing automated refresh
      not_documented otel

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
    • Cloudflare AI Gateway tracing note automated refresh
      `n.a.` — no OpenTelemetry or agent-trace representation documented for AI Gateway ([Logging](https://developers.cloudflare.com/ai-gateway/observability/logging/)) Configurable native OTLP export in JSON or Protobuf. Documented attributes include model, provider, usage, cost, custom metadata and prompt/completion payloads. Review content handling and collector access before enabling exports; validate parent-context propagation in your application.

      Local snapshot import; verify against the linked source before accepting.

      Source for this change ↗
  12. 9 changes

  13. 17 changes

    • Helicone Zero data retention manual edit
      Yes Not published

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Fireworks AI Zero data retention manual edit
      Not published Yes

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Fireworks AI zdr note manual edit
      Zero data retention is the default posture, not an upgrade. Zero retention is the default for open-model inference without opt-in logging. Responses API storage is on by default for 30 days; set store=False to disable it.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI zdr status manual edit
      yes depends

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI zdr note manual edit
      On by default and user-controlled, but note the boundary: ZDR applies only from the moment you enable it and does nothing about data already processed. ZDR must be enabled by disabling prompt/response storage in organization privacy settings. It applies prospectively and disables passthrough models.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI Logging default manual edit
      none full_content

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI logs note manual edit
      Inputs and outputs are not stored by default. Temporary caching may be used to improve performance unless configured otherwise. Prompts and responses are stored by default for product improvements. Organization administrators can disable storage to enable ZDR; training is a separate opt-in.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI Default content retention (days) manual edit
      0 Not published

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI retention note manual edit
      Zero by default. No retention window is published for the opt-in storage path. Prompt storage is on by default; a retention duration is not published on the privacy page. Enabling organization ZDR stops persistence of future request content.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI log retention note manual edit
      No prompt retention by default; endpoint metrics/events retention is not stated ([Privacy and security](https://docs.together.ai/docs/privacy-and-security), [Monitor endpoints and deployments](https://docs.together.ai/docs/dedicated-endpoints/monitoring)) Default prompt storage has no published duration on the reviewed privacy page. Organization ZDR disables future content storage; endpoint event retention is not specified. [Privacy and security](https://docs.together.ai/docs/privacy-and-security).

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Portkey Default content retention (days) manual edit
      90 Not published

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Portkey retention note manual edit
      Ninety days for logs and 365 for metrics by default. Self-hosted puts logs in your own S3-compatible store under your lifecycle policy. Plan and deployment dependent: Developer logs 3 days; Production logs 30 days ([Logs](https://portkey.ai/docs/product/observability/logs)). SaaS Enterprise logs 90 days and metrics 365 days; Hybrid logs follow customer storage lifecycle rules ([Security](https://portkey.ai/docs/enterprise/security)). No single default applies to all plans.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Portkey log retention note manual edit
      Developer 3 days (10k logs/month), Production 30 days (100k logs/month, $9 per additional 100k), Enterprise unlimited logs with retention unspecified ([Logs](https://docs.portkey.ai/docs/product/observability/logs)) Plan and deployment dependent: Developer logs 3 days; Production logs 30 days ([Logs](https://portkey.ai/docs/product/observability/logs)). SaaS Enterprise logs 90 days and metrics 365 days; Hybrid logs follow customer storage lifecycle rules ([Security](https://portkey.ai/docs/enterprise/security)). No single default applies to all plans.

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Amazon Bedrock Distributed tracing manual edit
      otel not_documented

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Together AI Distributed tracing manual edit
      otel not_documented

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Google Vertex AI Distributed tracing manual edit
      otel otel_partial

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
    • Groq Distributed tracing manual edit
      otel otel_integration

      Owner-authorized high-priority audit correction; source reviewed 2026-09-05.

      Source for this change ↗
  14. 1 change

    • Hugging Face Inference Providers catalog entry manual edit

      New catalog entry: zero-markup routing layer from Hugging Face (proprietary service, base URL router.huggingface.co/v1). Fans a single HF token out to 18 third-party inference partners, 15 of which serve chat completion, with automatic failover and :fastest / :cheapest / :preferred routing policies. Hugging Face charges the same rates as the provider with no additional fees. Scoped to the Inference Providers product only: the Hub and Inference Endpoints are out of scope, and Hub-level SOC 2, GDPR and EU-residency statements are deliberately not carried over to the gateway fields.

      Source for this change ↗
  15. 7 changes

    • Envoy AI Gateway catalog entry manual edit

      New catalog entry: Kubernetes-native OSS AI gateway from the Envoy Gateway project (Apache-2.0, self-host, Helm/K8s, `POST /v1/chat/completions` fully supported, v1.1.0 released 2026-08-21).

      Source for this change ↗
    • Higress catalog entry manual edit

      New catalog entry: CNCF Sandbox AI-native API gateway on Istio/Envoy with Wasm plugins (Apache-2.0, self-host + commercial Alibaba Cloud edition, 31 providers via AI Proxy plugin, v2.2.4 released 2026-08-13).

      Source for this change ↗
    • agentgateway catalog entry manual edit

      New catalog entry: Linux Foundation Rust-based data plane routing HTTP, gRPC, LLM, MCP and A2A traffic (Apache-2.0, self-host, drop-in OpenAI-compatible, v1.5.0 released 2026-08-27). Solo Enterprise for agentgateway is the same project as a commercial distribution.

      Source for this change ↗
    • Eden AI catalog entry manual edit

      New catalog entry: Managed multi-modality aggregator (Proprietary, 500+ models across 50+ providers, `POST /v3/chat/completions` at api.edenai.run, 5.5% platform fee at checkout with no markup on provider pricing). Not LLM-routing-first: half the product is non-LLM expert models (OCR, speech, document parsing).

      Source for this change ↗
    • Merge Gateway catalog entry manual edit

      New catalog entry: Managed LLM gateway from Merge (Proprietary, launched 31 March 2026, base URL api-gateway.merge.dev/v1). Routes to 24 model providers across 274 models, priced as LLM cost plus a 5% fee on the Pro plan. Distinct product line from Merge Unified; compliance material scoped to Unified is not being carried over.

      Source for this change ↗
    • MLflow AI Gateway catalog entry manual edit

      New catalog entry: LLM gateway feature inside MLflow Tracking Server (Apache-2.0, Linux Foundation, MLflow 3.15.2 released 2026-08-26). Correcting the deprecation myth: renamed to Deployments Server in 2.9.0, then explicitly reverted in 2.17.0 (2024-10-11); MLflow 3.0 removed the standalone deployment-server app but the gateway continues inside Tracking Server.

      Source for this change ↗
    • New API catalog entry manual edit

      New catalog entry: Self-host-only LLM key-management gateway (AGPL-3.0, hard-fork derivative of One API, 47,090 GitHub stars, v1.0.0-rc.30 released 2026-08-31). Highest star count of any candidate. Two GitHub advisories on record: CVE-2026-71479 (quota-overflow self-crediting) and CVE-2026-64859 (root token leak), both patched.

      Source for this change ↗
  16. 3 changes

    • Bifrost Models available manual edit
      1000 Not published

      Correction to our own reading, not a vendor change. We could not trace 1,000 to any page Bifrost publishes. The overview states "20+ AI providers" and names example models rather than a catalogue, so provider breadth is the only figure the vendor stands behind.

      Source for this change ↗
    • LiteLLM Models available manual edit
      1892 Not published

      Correction to our own reading, not a vendor change. We could not trace 1,892 to any page LiteLLM publishes. The providers documentation states no aggregate total and expresses coverage per provider; the repository says "100+ LLMs" and, confusingly, "100+ LLM providers" for the same figure. The field is now blank, which on this site reads "not published" rather than "no", and provider breadth is used for comparisons instead.

      Source for this change ↗
    • Together AI Models available manual edit
      19 110-200

      Correction to our own reading, not a vendor change. The earlier figure of 19 was counted from the example models named on the docs overview page, which is an illustrative list rather than the catalogue. Together publishes "200+ models" in its model directory and enumerates 110 on serverless; the range now carries both, because serverless is a subset of what the platform will run.

      Source for this change ↗
  17. 20 changes

    • Vercel AI Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • TrueFoundry AI Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Together AI catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Requesty catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Portkey catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Orq.ai Router catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • OpenRouter catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • LLM Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • LiteLLM catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Kong AI Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Helicone catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Groq catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Google Vertex AI catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Fireworks AI catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Cloudflare AI Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Braintrust Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Bifrost catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Azure AI Foundry catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Apache APISIX AI Gateway catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
    • Amazon Bedrock catalog entry manual edit

      Initial research pass. Date is the earliest per-field verification recorded for this product; the time of day is an ordering key, not evidence.

      Source for this change ↗
Back to top ↑

When each product was last verified

Each product's date is its oldest field check, not its newest — a single stale figure is what will mislead you, so that is what gets reported. Sourced fields counts how many values carry a citation: a low number means thin coverage, not that the product lacks the feature.

Product Type Sourced fields Oldest check Age
Apache APISIX AI Gateway Open source 29 73d
Google Vertex AI Cloud platform 32 73d
Kong AI Gateway Managed gateway 34 73d
LLM Gateway Open source 34 73d
Groq Inference provider 34 73d
Amazon Bedrock Cloud platform 35 73d
Azure AI Foundry Cloud platform 35 73d
Cloudflare AI Gateway Cloud platform 36 73d
Fireworks AI Inference provider 36 73d
TrueFoundry AI Gateway Cloud platform 37 73d
Braintrust Gateway Managed gateway 40 73d
Bifrost Open source 40 73d
Together AI Inference provider 40 73d
LiteLLM Open source 44 73d
Portkey Managed gateway 45 73d
Helicone Managed gateway 47 73d
Orq.ai Router Managed gateway 47 73d
Requesty Managed marketplace 48 73d
OpenRouter Managed marketplace 49 73d
Vercel AI Gateway Managed gateway 52 73d
New API Open source 44 69d
agentgateway Open source 45 69d
MLflow AI Gateway Open source 45 69d
Higress Open source 47 69d
Eden AI Managed marketplace 54 69d
Agent Router Open source 66 69d
Merge Gateway Managed gateway 130 69d
Hugging Face Inference Providers Managed marketplace 88 68d
Respan Managed gateway 176 56d
AI Gateway HQ Managed gateway 170 54d
Velokey Managed marketplace 100 52d
SteadyGateway Managed marketplace 145 21d
Back to top ↑