RepoRadar
Leave a Tip
Item detail
github.com

ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs)

ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs) is an AI project that RepoRadar is tracking in its Radar section, currently rated Gold tier with a 'try now' verdict. Its strongest signal is workflow potential, scored 10.0 out of 10.

Open on GitHub →Back to radar
Score9.0
Popularity0.0
Risklow
TierGold
Score breakdown
Usefulness9.0
Novelty9.0
Momentum8.0
Maturity7.0
Open-source/build8.4
Evidence7.2
Workflow potential10.0
Setup ease6.4

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Most coding-agent developers today who want their agent to plug into a code editor (Zed / VS Code / JetBrains) have been either (a) writing custom JSON-RPC bridges per editor (no canonical protocol), (b) reaching for non-standard integrations like stdin / stdout file-watch that require custom agent-side code, or (c) maintaining separate transport layers per editor.

Who should use it

Coding-agent developers building new agents that need editor integrations + code-editor developers wanting to embed a coding agent in their product without building a custom protocol bridge + AI agent developers standardizing on ACP for their agent / client integration + SDK / framework authors building downstream ACP clients or agents + anyone integrating coding-agent capabilities into a code editor (Zed / VS Code / JetBrains / etc.) + any developer wanting a clean Apache-2.0 open standard for Editor/agent developers + 5-SDK users that want the 5 first-party SDKs (Rust `agent-client-protocol` + `agent-client-protocol-schema` crates on crates.io; Kotlin `acp-kotlin` for JVM; Java `java-sdk`; Python `python-sdk`; TypeScript `@agentclientprotocol/sdk` on npm) -- the right 5-SDK primitive for any developer who has been writing custom transport-layer code per language Editor/agent developers + JSON-Schema-v1-v2 users that want the JSON Schema artifacts at `schema/v1/schema.json` + `schema/v2/schema.json` attached to GitHub `schema-v*` releases -- the right JSON-Schema primitive for any SDK generator author who needs stable schema structure for code generation Editor/agent developers + protocolVersion-negotiation users that want the `protocolVersion` field in `initialize` for wire-compat negotiation -- the right protocolVersion primitive for any protocol author who has been fighting schema-vs-wire compat drift Editor/agent developers + agentclientprotocol-com-adoption-directories users that want the Agents + Clients + official Libraries + Community Libraries directories on `agentclientprotocol.com` -- the right adoption-directories primitive for any developer who has been searching for an editor/agent pair that supports a canonical protocol Editor/agent developers + JSON-RPC-envelope users that want the JSON-RPC envelope over stdio / WebSocket -- the right JSON-RPC-envelope primitive for any developer who has been writing custom request/response wrappers per transport

Who should skip it

Consider ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs) lower priority if you already have a working solution in this category.

About this signal

ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs) is tracked by RepoRadar as an AI project in the Radar section. First seen 2026-07-08; the source record was last checked on 2026-07-08. The current verdict is 'try now' with a Gold tier and moderate setup difficulty. Across RepoRadar's eight signals, ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs) is strongest on workflow potential (10.0) and practical usefulness (9.0) and weakest on setup ease (6.4) — a profile worth weighing against your own priorities. This page summarizes the public evidence on the linked source page and states where additional review is still needed.

How this item is evaluated

The ACP: Apache-2.0 Open Standard (Current Stable v1) for Connecting Any Code Editor to Any Coding Agent (JSON-RPC, 5 SDKs) record combines a 9.0/10 composite score with separate popularity (0.0), risk (low), and setup (moderate) signals. See the scoring methodology for the current weights and evidence definitions.

Putting this into practice? Read How to vet an AI agent or MCP server before you wire it in for the checklist behind this score.

Risk explanation

The 3; 600* / 296-fork / 16-subscriber repo is at active maintenance with the protocol v1 stable but the consumer SHOULD note the project is a protocol + SDK bundle (not a single binary); so adoption requires integrating one of the 5 SDKs (Rust / Kotlin / Java / Python / TypeScript) into an existing editor or agent -- not a drop-in install; the consumer SHOULD note the protocol is JSON-RPC over stdio / WebSocket.

Evidence links
Closest alternatives / related signals
open-source apache-2-0 agentclientprotocol agent-client-protocol acp open-standard protocol-version-1 json-rpc
Verification record

What RepoRadar actually verified

Tested in a bounded workflow

Bounded representative workflow retained by RepoRadar verification harness. Last checked 2026-07-14T04:56:48.478436Z.

failed · cohort-20260713-acp-ndjson-read-callback-workflow

Tester
RepoRadar automated local verification harness
Started
2026-07-14T04:55:39.053616Z
Completed
2026-07-14T04:55:43.625035Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
1 minute(s)
Evidence scope
Bounded setup or capability check
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Created 3 synthetic fixture file(s) inside the disposable work directory; retained hashes prove the exact inputs.
  • Installed the pinned official ACP TypeScript SDK with lifecycle scripts disabled, then connected client and agent applications through paired NDJSON streams.
  • Negotiated protocol v1 and read-text-file capability, created a fixture session, issued a prompt, and serviced exactly one read callback for the allowlisted marker file.
  • Collected the agent message notification and end-turn response, then retained the negotiated version, callback count, update kind/text, and stop reason.
  • Executed bounded check: Connect an ACP client and agent over the SDK's NDJSON transport, negotiate capabilities, create a session, and service one client-side text-file read callback.
  • Captured the complete sanitized stdout, stderr, exit status, artifact checks, and 4.57-second wall time.
Observed results
  • Check exited 1 after 4.57 seconds; failure output is retained and the verification stage must not advance.
Observed strengths
  • No product strength was established in this failed attempt; diagnostic evidence is retained.
Friction
  • The protocol requires explicit bidirectional stream wiring and capability declarations even for a one-file read; production integrations must additionally handle permissions, cancellation, and connection loss.
  • Setup or runtime emitted 17 stderr line(s); the complete warnings/errors are preserved in the retained log.
Limitations
  • The in-process deterministic agent validates ACP framing, initialization, session, prompt, notification, and read-file callback behavior; it does not test a production editor, model-backed agent, subprocess crash recovery, permission UI, or remote transport.
  • This credential-free disposable workflow does not establish operator use, production scale, model quality, reliability under sustained use, or team adoption.

Pricing assessment: The official Apache-2.0 SDK ran locally; npm package retrieval was the only external dependency and no account, paid service, or model was used.

Privacy assessment: Only the synthetic marker crossed the in-process NDJSON streams; dependency installation contacted npm, while no prompt or file content left the disposable process.

Open retained test log →

failed · cohort-20260713-acp-ndjson-read-callback-workflow-v2

Tester
RepoRadar automated local verification harness
Started
2026-07-14T04:56:21.804627Z
Completed
2026-07-14T04:56:24.751656Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
1 minute(s)
Evidence scope
Bounded setup or capability check
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Created 3 synthetic fixture file(s) inside the disposable work directory; retained hashes prove the exact inputs.
  • Installed the pinned official ACP TypeScript SDK with lifecycle scripts disabled, then connected client and agent applications through paired NDJSON streams.
  • Negotiated protocol v1 and read-text-file capability, created a fixture session, issued a prompt, and serviced exactly one read callback for the allowlisted marker file.
  • Collected the agent message notification and end-turn response, then retained the negotiated version, callback count, update kind/text, and stop reason.
  • Executed bounded check: Connect an ACP client and agent over the SDK's NDJSON transport, negotiate capabilities, create a session, and service one client-side text-file read callback.
  • Captured the complete sanitized stdout, stderr, exit status, artifact checks, and 2.95-second wall time.
Observed results
  • Check exited 13 after 2.95 seconds; failure output is retained and the verification stage must not advance.
Observed strengths
  • No product strength was established in this failed attempt; diagnostic evidence is retained.
Friction
  • The protocol requires explicit bidirectional stream wiring and capability declarations even for a one-file read; production integrations must additionally handle permissions, cancellation, and connection loss.
  • Setup or runtime emitted 8 stderr line(s); the complete warnings/errors are preserved in the retained log.
Limitations
  • The in-process deterministic agent validates ACP framing, initialization, session, prompt, notification, and read-file callback behavior; it does not test a production editor, model-backed agent, subprocess crash recovery, permission UI, or remote transport.
  • This credential-free disposable workflow does not establish operator use, production scale, model quality, reliability under sustained use, or team adoption.

Pricing assessment: The official Apache-2.0 SDK ran locally; npm package retrieval was the only external dependency and no account, paid service, or model was used.

Privacy assessment: Only the synthetic marker crossed the in-process NDJSON streams; dependency installation contacted npm, while no prompt or file content left the disposable process.

Open retained test log →

passed · cohort-20260713-acp-ndjson-read-callback-workflow-v3

Tester
RepoRadar automated local verification harness
Started
2026-07-14T04:56:45.313686Z
Completed
2026-07-14T04:56:48.478436Z
Environment
Windows 10 AMD64; Python 3.11.9; credential-stripped child environment; disposable home/cache
Install/setup time
1 minute(s)
Evidence scope
Bounded representative workflow
Cleanup
Per-check temporary home and work directory removed. Shared cohort package cache removed.
Actions exercised
  • Created a disposable home, work directory, and isolated package cache with credential-like environment variables excluded.
  • Created 3 synthetic fixture file(s) inside the disposable work directory; retained hashes prove the exact inputs.
  • Installed the pinned official ACP TypeScript SDK with lifecycle scripts disabled, then connected client and agent applications through paired NDJSON streams.
  • Negotiated protocol v1 and read-text-file capability, created a fixture session, issued a prompt, and serviced exactly one read callback for the allowlisted marker file.
  • Collected the agent message notification and end-turn response, then retained the negotiated version, callback count, update kind/text, and stop reason.
  • Executed bounded check: Connect an ACP client and agent over the SDK's NDJSON transport, negotiate capabilities, create a session, and service one client-side text-file read callback.
  • Captured the complete sanitized stdout, stderr, exit status, artifact checks, and 3.16-second wall time.
Observed results
  • Command exited 0 after 3.16 seconds.
  • The ACP SDK completed protocol negotiation, session creation, prompt routing, one client file callback, one streamed update, and a normal end turn over NDJSON framing.
  • Expected marker 'CHECK_OK protocol=1 session=session-fixture reads=1 updates=1 stop=end_turn' was observed in retained output.
  • Validated result.json: 5 required marker(s) present and 0 excluded marker(s) absent; size and SHA-256 are retained.
Observed strengths
  • The official client and agent builders exercised both request directions and the session update path over the SDK's serialized transport without credentials or an external agent.
Friction
  • The protocol requires explicit bidirectional stream wiring and capability declarations even for a one-file read; production integrations must additionally handle permissions, cancellation, and connection loss.
  • Setup or runtime emitted 2 stderr line(s); the complete warnings/errors are preserved in the retained log.
Limitations
  • The in-process deterministic agent validates ACP framing, initialization, session, prompt, notification, and read-file callback behavior; it does not test a production editor, model-backed agent, subprocess crash recovery, permission UI, or remote transport.
  • This credential-free disposable workflow does not establish operator use, production scale, model quality, reliability under sustained use, or team adoption.

Pricing assessment: The official Apache-2.0 SDK ran locally; npm package retrieval was the only external dependency and no account, paid service, or model was used.

Privacy assessment: Only the synthetic marker crossed the in-process NDJSON streams; dependency installation contacted npm, while no prompt or file content left the disposable process.

Open retained test log →

Verification sources

Longitudinal intelligence

How this decision record is moving

Raw history JSON →

29 dated snapshots retained from 2026-07-08 through 2026-08-13; see the snapshot index for explicit coverage gaps. Stars, version, release, pricing, integration, risk, maintenance, verdict, score, and momentum fields remain explicit even when a source has not reported them. Repository momentum is a normalized 0–10 RepoRadar signal; GitHub stars appear only where the popularity monitor retained exact timestamped observations.

RepoRadar score9.0 current · +0.0 net
Repository momentum9.3 current · +1.3 net
GitHub stars (observed)3,966 current · +314 net
GitHub stars3,966 exact observation
Versionschema-v2.0.0-alpha.2
Last release2026-07-21T15:09:38Z
Maintenanceactive
Current risklow
Current verdicttry now
Pricing baselineNo structured commercial pricing baseline
Pricing checkedNot applicable or not recorded
Pricing freshnessNo dated commercial pricing review
Integrations baselineNo structured integrations recorded

Recent dated points

DateScoreMomentumStarsRiskVerdictMaintenance
2026-08-139.09.33,966lowtry nowactive
2026-08-129.09.63,957lowtry nowactive
2026-08-119.09.63,933lowtry nowactive
2026-08-109.09.33,917lowtry nowactive
2026-08-099.09.33,913lowtry nowactive
2026-08-089.09.63,906lowtry nowactive
2026-08-079.09.33,847lowtry nowactive
2026-08-069.08.0Not recordedlowtry nownot recorded
2026-08-059.08.0Not recordedlowtry nownot recorded
2026-08-049.09.33,847lowtry nowactive
2026-08-039.09.33,847lowtry nowactive
2026-08-029.09.33,840lowtry nowactive

Why the record changed

stars changed

Stars changed: 3957 → 3966.

stars changed

Stars changed: 3933 → 3957.

stars changed

Stars changed: 3917 → 3933.

stars changed

Stars changed: 3913 → 3917.

stars changed

Stars changed: 3906 → 3913.

stars changed

Stars changed: 3847 → 3906.

stars changed

Stars changed: 3840 → 3847.

stars changed

Stars changed: 3832 → 3840.

stars changed

Source-observed stars changed: 3827 → 3832. This reports the retained observation delta and does not infer why the upstream change occurred.

stars changed

Source-observed stars changed: 3825 → 3827. This reports the retained observation delta and does not infer why the upstream change occurred.

stars changed

Stars changed: 3787 → 3804.

stars changed

Stars changed: 3676 → 3787.