Arch is live · free · last verified 22 Oct 2026 · viability healthy
Arch
Plano is an open-source AI-native proxy for smart LLM routing, agent orchestration, and guardrails in agentic apps.
Plano earns its place when agent middleware has become copy-paste across services — YAML agent definitions, Filter Chains, and zero-code Agentic Signals remove real toil, and you keep your stack. The catch is operational: production means self-hosting the Plano routing LLMs, and the project is young enough that you'll be reading source and filing issues. Worth a pilot on a non-critical agent before you commit.
Verdict generatedVerified 7d ago · liveness 72/100 · cite: rightaichoice.com/tools/arch
- Developers building multi-agent systems who need orchestration plus smart LLM routing
- Platform teams self-hosting agent middleware on their own infrastructure
- Teams wanting unified guardrails, moderation, and observability without touching agent code
- Organizations seeking an open-source, vendor-neutral alternative to managed AI gateways
- Non-technical users looking for a no-code agent platform
- Teams requiring a fully managed cloud service with SLAs and dedicated support
- Developers who need prebuilt third-party connectors out of the box
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Plano if you need a managed agent platform with an SLA rather than an Envoy-based proxy you deploy and operate yourself, or if you want prebuilt third-party connectors out of the box.
Running the Plano routing LLMs in production means provisioning your own GPU or CPU capacity — the free US-central hosted orchestrator is a first-run experience, not a production endpoint.
Plano is free under Apache-2.0 with no published tiers, which puts it below managed AI gateway subscriptions and per-seat agent platforms. The real cost comparison is against your own time: teams without a platform engineer should price a managed gateway or framework middleware instead, while larger platform organizations that already run Envoy and Kubernetes will find self-hosted Plano cheaper than per-request gateway fees.
In short
Arch — Plano is an open-source AI-native proxy for smart LLM routing, agent orchestration, and guardrails in agentic apps. Best for Developers building multi-agent systems who need orchestration plus smart LLM routing, Platform teams self-hosting agent middleware on their own infrastructure, Teams wanting unified guardrails, moderation, and observability without touching agent code. Free to use.
What is Arch?
Arch — Plano is an open-source AI-native proxy for smart LLM routing, agent orchestration, and guardrails in agentic apps. Best for Developers building multi-agent systems who need orchestration plus smart LLM routing, Platform teams self-hosting agent middleware on their own infrastructure, Teams wanting unified guardrails, moderation, and observability without touching agent code. Free to use.
Plano is an Apache-2.0 proxy server and data plane for agentic applications, built on Envoy by its contributors and living at katanemo/plano. It's aimed at production-focused developers who are tired of rebuilding routing logic, guardrail hooks, and observability glue in every microservice. You declare agents in YAML — an id, a URL, and a natural-language description of what each one handles — and Plano runs the orchestration plumbing out of process. The centerpiece is smart LLM routing. Plano routes by model name, by semantic alias, or automatically via preferences, and its default router, plano_orchestrator_v1, is a 4B-parameter model that reads your agent descriptions and dispatches requests. Because agents are just HTTP servers behind an OpenAI-compatible chat completions endpoint, you keep whatever language or AI framework you already use. The rest of the bundle is the "hidden middleware" most teams hand-roll: zero-code Agentic Signals capture with OpenTelemetry traces and metrics, random-sampling tracing for evaluation, and Filter Chains for jailbreak protection, moderation, and memory hooks enforced consistently across agents. Deployment is a sidecar or standalone service via Docker, and you can configure edge proxying, LLM routing, or both together. Plano's family of LLMs is hosted free in US-central for a first-run developer experience; for production scale you run the models locally or reach out on Discord for API keys. Against managed AI gateways and framework-bundled middleware, Plano is the vendor-neutral self-hosted pick for teams that want routing and guardrails to live outside application code — at the cost of running the routing models yourself.
Key features
- AI-native proxy server and data plane built on Envoy
- Low-latency orchestration and routing between multiple agents
- Add new agents via YAML without modifying app code
- Smart LLM routing by model name, alias, or preferences
- plano_orchestrator_v1 4B-parameter routing model
- Zero-code Agentic Signals capture across every agent
- OpenTelemetry traces and metrics with no instrumentation
- Random-sampling tracing for evaluation
- Filter Chains for jailbreak protection and moderation
- Memory hooks enforced consistently across agents
- OpenAI-compatible chat completions endpoint
- Agents are plain HTTP servers in any language
- Sidecar or standalone deployment with Docker
- Configurable edge proxying, LLM routing, or both
- Free hosted Plano LLMs in US-central for first run
Behind the verdict
We'd reach for Plano the moment routing logic starts duplicating across services. YAML agent declarations plus natural-language descriptions mean a new agent is a config entry, not a code change, and the OpenAI-compatible endpoint lets your Python, TypeScript, or Go agents stay untouched. When to pick it: multi-agent systems where you need orchestration, smart LLM routing, and guardrails in one data plane, and you own your infrastructure. Platform teams who want vendor-neutral middleware under Apache-2.0 get exactly that, and the Filter Chains plus Agentic Signals cover the safety and observability work you'd otherwise wire per service. When to pass: non-technical teams wanting a no-code agent platform, or anyone who needs a fully managed service with an SLA. The hosted Plano LLMs are free in US-central for first runs, but production scale means running the models locally — that's real infrastructure work, and for some teams a managed AI gateway is the saner trade. Closest alternative: framework-bundled middleware, or managed AI gateways. Plano's pitch is that it lives outside application code instead of inside it. In practice the caveats are maturity and support — expect to read source and file issues when something misbehaves. One practical note: random-sampling tracing makes evaluation cheap to switch on, so start there before you commit to a router model choice.
Models under the hood
as of 2026-10-25
Limitations
- Plano is self-hosted, requiring DevOps effort for deployment and maintenance.
- It does not yet have an extensive library of pre-built integrations, so custom connectors may be needed.
- The project is relatively new (roughly 7.1k stars, 753 commits), so documentation and community resources are still maturing.
- The free hosted Plano-Orchestrator runs in US-central as a first-run developer experience only; scaling to production means running the models locally or contacting the team on Discord for API keys.
- The demo orchestration model, plano_orchestrator_v1, is a 4B-parameter router, so you should benchmark it against your own traffic before depending on it.
as of 2026-10-12
Researching Arch? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Arch still active?
Arch is live · free · last verified 22 Oct 2026 · viability healthy.
Viability score
How well maintained and how widely used is Arch? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What do users say about Arch?
What people actually say about Arch — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
Researched26 mentions across 5 sources (Hacker News, Product Hunt, App Store, Stack Overflow, Lemmy), 56 more we could not attribute · researched Oct 22, 2026.
Weighted by the 82 posts each of 5 sources contributed.
- +Envoy-based proxy runs as a sidecar or standalone, fitting existing service-mesh topologies
- +OpenAI-compatible chat completions endpoint means agents can be plain HTTP servers in any language
- +YAML-declared agents with natural-language descriptions remove routing code from application services
- +Vendor-neutral, self-hosted option avoids lock-in to managed AI gateways like Portkey or Cloudflare AI Gateway
- +Zero-code Agentic Signals plus OpenTelemetry traces and metrics with no instrumentation
- −No verified community feedback in this dataset — every scraped 'Arch' post is about something else
- −Rebrand from Arch to Plano splits discoverability and muddles search results badly
- −Free hosted Plano LLMs are US-central only and described as a first-run experience
- −Scaling production means self-hosting a 4B routing model or negotiating keys on Discord
- −Operational ownership covers Envoy, filter chains, YAML config, and telemetry pipelines
- • GPU or cloud capacity to self-host the 4B routing model at production scale
- • AI inference spend on whichever upstream model providers the router calls
- • Salaried engineering time for Envoy operations, filter chain maintenance, and OTel backend costs
- • No stated support contract — procurement teams may need to negotiate separately
How we verified this
Every listed fact about Arch— pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 22 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified Arch 12 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed pricing page
- — changed our verdict, who should skip it, who it suits, summary, description, features, pricing tiers
- — changed pricing tiers, summary, description, our verdict, pricing page
- — changed who it suits, description, use cases, our verdict, summary, who should skip it, hidden costs, setup time, limitations, switching in, switching out, pricing fit
- — changed pricing page
- — changed our verdict, pricing tiers, summary, features, who it suits, who should skip it, description
Showing the 6 most recent of 12 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Arch cost in 2026?
Arch: Free to use. Pricing last verified 22 Oct 2026.
Plans compared
Every published Arch tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Open Source (Apache-2.0) | Free | — | Platform teams and backend developers who are willing to operate an Envoy-based proxy and self-host the routing LLMs.Adds: Starting tier and the only published tier: the full proxy server and data plane under Apache-2.0 at $0, plus a free hosted Plano-Orchestrator in US-central for first-run evaluation. |
| Hosted Plano LLMs (US-central) | Free | — | — |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where Arch's pricing actually pencils out — and where peers do it cheaper.
Plano is free under Apache-2.0 with no published tiers, which puts it below managed AI gateway subscriptions and per-seat agent platforms. The real cost comparison is against your own time: teams without a platform engineer should price a managed gateway or framework middleware instead, while larger platform organizations that already run Envoy and Kubernetes will find self-hosted Plano cheaper than per-request gateway fees.
Has Arch's pricing changed?
No plan price for Arch has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Arch?
Pick Arch if you are Developers building multi-agent systems who need orchestration plus smart LLM routing, Platform teams self-hosting agent middleware on their own infrastructure, Teams wanting unified guardrails, moderation, and observability without touching agent code, Organizations seeking an open-source, vendor-neutral alternative to managed AI gateways.
Look elsewhere if you are Non-technical users looking for a no-code agent platform, Teams requiring a fully managed cloud service with SLAs and dedicated support, Developers who need prebuilt third-party connectors out of the box, Teams unwilling to self-host the Plano routing LLMs for production scale.
Skip it if Skip Plano if you need a managed agent platform with an SLA rather than an Envoy-based proxy you deploy and operate yourself, or if you want prebuilt third-party connectors out of the box.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Arch actually fits — and what changes day-one when you adopt it.
You have three services, each calling OpenAI directly with hard-coded model names and its own retry logic. You deploy Plano as a sidecar, point every service at its OpenAI-compatible chat completions endpoint, and move model choice into YAML aliases so switching models no longer means a redeploy.
Outcome: Model changes become config changes, and every service shares one routing policy instead of three divergent ones.
Several teams ship agents in Python, Go, and TypeScript. You register each agent's URL and a natural-language description in Plano's YAML, let the orchestrator route user requests, and attach a Filter Chain for jailbreak protection and moderation plus memory hooks that apply to all of them.
Outcome: One place to change guardrails and routing, and OTEL traces plus Agentic Signals land in your existing observability stack without per-service instrumentation.
You run the travel-agent demo (weather and flight agents through plano_orchestrator_v1) with random-sampling tracing switched on, then inspect Agentic Signals to see where routing picks the wrong agent.
Outcome: You get an evidence-based read on whether the 4B orchestrator handles your traffic before committing to a self-hosted production deployment.
Use cases
- Build a personalized customer support agent that securely accesses internal APIs and databases with routing handled outside your app code.
- Add safety guardrails and OpenTelemetry logging to an existing LLM-powered chatbot without modifying app code.
- Deploy a multi-model agent that routes requests to the best LLM based on cost, latency, or alias preference.
- Monitor and trace agent decisions in real time via Agentic Signals and OTEL metrics to debug and improve performance.
- Prototype an agentic workflow that integrates with multiple third-party APIs using natural-language agent descriptions.
- Centralize jailbreak protection and moderation policy across a fleet of agents via Filter Chains instead of per-service hooks.
- Run a sidecar next to each service so model routing and guardrails are consistent across languages and frameworks.
- Evaluate orchestration quality with random-sampling tracing on the travel-agent demo before rolling out to production.
Setup time & first value
How long it actually takes to get something useful out of Arch — broken out by persona, not the marketing-page minute.
Developer on a laptop: minutes to run the Docker-based quickstart against the free US-central hosted Plano-Orchestrator. Production self-host: hours to days, dominated by provisioning capacity for the Plano LLMs, wiring the proxy into your Kubernetes or sidecar topology, and connecting OTEL to your observability backend.
Switching to or from Arch
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From hard-coded LLM calls in each service: point services at Plano's OpenAI-compatible endpoint, then move model names into YAML aliases and preferences.
- →From an in-app framework middleware layer: lift routing, guardrails, and tracing into Plano Filter Chains and delete the per-service copies.
- →From a managed AI gateway: reimplement your route rules as Plano YAML agents and routes, then run the proxy as a sidecar or standalone service.
- ↗To a managed AI gateway: replicate your YAML routes and Filter Chains in the gateway's config, then move the OpenAI-compatible endpoint URL in each service.
- ↗To framework-bundled middleware: pull guardrails and routing back into application code, accepting the loss of cross-language centralization.
- ↗To a direct model provider SDK: remove the Plano endpoint, restore per-service model selection, and re-add observability instrumentation in each service.
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Arch”, and we withheld 6: 6 could not be judged, because “Arch” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Arch.
Official links
What are the best Arch alternatives?
The closest alternatives to Arch we list are Plano, TrueFoundry AI Gateway and 9router. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Plano | Freemium | Pricing not published | — |
| TrueFoundry AI Gateway | Freemium | Pricing not published | — |
| 9router | Freemium | Pricing not published | — |
Alternatives to Arch
View all
Plano
Plano is an AI-native proxy and data plane handling agent routing, orchestration, traces, and guardrails for agentic apps.
TrueFoundry AI Gateway
One OpenAI-compatible gateway in front of 1,600+ models, with quotas, RBAC, guardrails and an MCP gateway for agent tool calls.
Tools that pair well with Arch
Common stack mates teams adopt alongside Arch, with the specific reason each pairing earns its keep.
Plano
Plano is an AI-native proxy and data plane handling agent routing, orchestration, traces, and guardrails for agentic apps.
TrueFoundry AI Gateway
One OpenAI-compatible gateway in front of 1,600+ models, with quotas, RBAC, guardrails and an MCP gateway for agent tool calls.
9router
Open-source AI routing gateway that fronts 60+ providers behind one OpenAI-compatible endpoint at localhost:20128/v1.
Arch vs its closest rival
Arch's closest head-to-head is with Spider Cloud — Spider Cloud and Arch solve entirely different problems: Spider Cloud pulls live web data into AI pipelines, while Arch orchestrates and secures agent-to-LLM communication. Pick Spider Cloud if your bottleneck is getting structured web content fast (news, product pages, search results). Pick Arch if you're wiring multiple agents together and want built-in moderation, tracing, and model routing without reinventing the wheel. They are complementary – you could use Spider Cloud as a web tool inside an Arch-routed agent. Verdict generated
Head-to-head comparisons
Arch vs Spider Cloud
Spider Cloud and Arch solve entirely different problems: Spider Cloud pulls live web data into AI pipelines, while Arch orchestrates and secures agent-to-LLM communication. Pick Spider Cloud if your bottleneck is getting structured web content fast (news, product pages, search results). Pick Arch if you're wiring multiple agents together and want built-in moderation, tracing, and model routing without reinventing the wheel. They are complementary – you could use Spider Cloud as a web tool inside an Arch-routed agent.
GeneratedArch vs Temporal Ai
If you need bulletproof durability for long-running AI agents or microservices that survive crashes and retries, choose Temporal. If you primarily need a lightweight, open-source proxy to orchestrate multiple agents with built-in safety and observability, Arch is the better fit. Temporal is more powerful for mission-critical workflows; Arch is simpler for multi-agent routing.
GeneratedArch vs Presto Voice
Presto Voice and Arch serve completely different needs. Presto Voice is a specialized drive-thru voice AI for QSR chains, delivering up to 95% automation and upselling boosts—ideal for franchise operators. Arch is an open-source AI proxy for developers building multi-agent systems, offering routing, safety, and observability. Choose based on whether you need to automate restaurant ordering or orchestrate agentic workflows.
GeneratedFrequently asked questions
Used Arch? Help shape our editorial sentiment research.