Repository navigation
Releases
Releases · Sagargupta16/bedrock-multi-model-mcp
v0.9.0
Added
bedrock_paneltool: 2-5 models discuss one prompt in turn and the tool returns the transcript. Unlikebedrock_compareandbedrock_judge, the models answer one after another and each sees what the earlier ones said.rolesgives each model one of six presets (critic,builder,fact-checker,devils-advocate,simplifier,moderator), free text up to 200 characters, ornull. Every model is told who else is on the panel and in which role.rounds(1 to 3, default 1) lets each model answer again having seen everything so far;summarizenames a model that closes with where the panel agrees and disagrees;level(cheap,mediumorintense) picks a default panel of three whenmodelsis absent.- A model's view is one user message holding the question and the turns so far, each prefixed with the speaker's name in square brackets, its own turns marked
(you). What a model said never reaches another model's system prompt. - A model that fails loses its turn and the panel carries on, as in
bedrock_compare. The same model may be listed twice only with different roles.max_tokensandeffortwork as inbedrock_compare. - The idea comes from the group-chat round in confab; no code or dependency was taken from it.
- GPT-6.2 Astra (
us.openai.gpt-6.2-astra) and GPT-6.2 Sol (us.openai.gpt-6.2-sol), on Bedrock since 2026-11-17. Both were live-probed on 2026-11-20: they takeschemaandeffortfromlowtomax, and rejecttemperature, so they are flaggednoTemperature. bedrock_doctortool: one small billed call (a one-word prompt, at most 32 output tokens) to every registry text model, or only themodelsyou pass, withconcurrency1 to 8 (default 4). The table gives one status per model:ok,throttled,denied,invalidorerror, with the latency and a short note.filesoption onbedrock_ask,bedrock_compare,bedrock_judgeandbedrock_panel: local text files are added to the prompt as fenced blocks. Files are read only from underBEDROCK_MCP_FILES_ROOT(default the server's working folder, refused when that is a home folder or a drive root), at most 10 files, 256 KB each and 1 MB in total; names that look like secret stores (.env,.npmrc,.git-credentials, private keys, anything under.sshor.awsand the like), network and device paths, binary files and anything outside the root are refused, and one refusal fails the call before any model is billed. The content is sent to the model provider.- Cost estimates: every text model in
src/data/text-models.jsonnow has an optionalprice(USD per 1M input and output tokens).bedrock_ask,bedrock_compare,bedrock_judgeandbedrock_panelshow an estimated cost, andbedrock_list_modelshas a price column. The prices are from the AWS Price List bulk files for us-east-1, read 2026-11-20, on-demand standard tier; for ids with theus.profile prefix the non-global price was used. Cache, batch and long-context pricing are not modeled, so a figure is an estimate. level(cheap,medium,intense) onbedrock_ask,bedrock_compareandbedrock_judge, with the same presets asbedrock_panel.bedrock_askuses the first model of the level;modelonbedrock_askandmodelsonbedrock_compareandbedrock_judgeare now optional whenlevelis given, andmodelswins overlevel.bedrock_askrejectsmodeltogether withlevel.- One-click
.mcpbbundle for Claude Desktop and other MCPB hosts:npm run pack:mcpb(scripts/pack-mcpb.mjs, manifest inmcpb/manifest.json) buildsbedrock-multi-model-mcp-<version>.mcpb, and abundlejob inrelease.ymlattaches it to the GitHub release after the publish job succeeds. - MCP prompts:
second-opinion,compare-modelsandpanel-discussion(ready-made instructions that callbedrock_ask,bedrock_compareandbedrock_panel). conversation_idonbedrock_ask: opt-in follow-up conversations. Earlier turns of the same id are sent along as context. Kept in process memory only, forgotten 30 minutes after the last turn or when the server restarts, at most 20 turns per conversation, 50 conversations and 200,000 characters per conversation; the id is 1 to 64 letters, digits,-or_.- New-model watcher:
npm run new-models(scripts/check-new-models.mjs) lists Bedrock model cards that the catalog neither has nor has deliberately skipped inscripts/model-skips.json. Anew-modelsjob inlifecycle.ymlruns it weekly and on manual runs, and opens or updates one issue. - Luma Ray 2 keyframes on
bedrock_generate_video:start_image_pathandend_image_pathturn local JPEG or PNG files (512x512 to 4096x4096 pixels) into the first and last frame. The images are checked before the job is started. - Image inputs on
bedrock_embed_similarity:imagestakes local JPEG or PNG paths to embed beside the texts, for the multimodal modelstitan-multimodalandcohere. The tool needs at least 2 inputs and at most 50 acrosstextsandimages.
Changed
- Breaking: bare
gpt-astranow resolves to GPT-6.2 Astra andgpt-solto GPT-6.2 Sol. Theintensepanel ofbedrock_panelusesgpt-astra, so it moves with it.
Removed
- Breaking: GPT-6 Astra (
gpt-6-astra) and GPT-6.1 Sol (gpt-6.1-sol). AWS's cards name GPT-6.2 Astra and GPT-6.2 Sol as their next generation.
Security
npm auditwent from 2 findings to 0, closing the 7 Dependabot alerts opened on 2026-11-13: hono 4.13.13 -> 4.13.19 (5 advisories) and @hono/node-server 2.1.5 -> 2.1.9 (2 advisories), both pulled in by@modelcontextprotocol/sdk. The stdio server never starts hono's HTTP server.@aws-sdk/client-bedrock-runtimefloor raised to ^3.1173.0.
v0.8.0
Added
- Claude Fable 5.2 (
us.anthropic.claude-fable-5-2, on Bedrock since 2026-11-09), GPT-6 Astra (gpt-astra, the model AWS's card calls OpenAI's most capable) and GPT-6 Luna (gpt-luna, OpenAI's fast tier). Each was live-probed on 2026-11-10: all three takeschemaandeffortfromlowtomax, and rejecttemperature, so they are flaggednoTemperature.
Changed
- Breaking: bare
fableandclaude-fablenow resolve to Fable 5.2. Fable 5.1 stays reachable asfable-5.1andclaude-fable-5.1. @aws-sdk/client-bedrock-runtimefloor raised to ^3.1165.0.
Removed
- Breaking: Claude Fable 5 (
fable-5), now two versions behind Fable 5.2.
Not added
- Kimi K3.1 Pro - Highspeed (
us.moonshotai.kimi-k3.1, on Bedrock since 2026-11-04): every probe call on 2026-11-10 was throttled ("Too many tokens"), on the US and the global profile, so it stays out until a call answers.kimistill resolves to Kimi K3.
v0.7.1
Changed
- README setup installs from npm:
claude mcp add ... -- npx -y bedrock-multi-model-mcpand annpxserver block for Claude Code and Claude Desktop, instead of cloning and building. Building from source moved to "Run from source". - First release through
release.ymlwith the npm trusted publisher (GitHub OIDC, no token), so this version carries npm provenance; 0.7.0 was published by hand.
v0.7.0
Added
bedrock_edit_imagetool for the 13 Stability AI Image Services on Bedrock (us.stability.*profiles, us-west-2):inpaint,outpaint,erase,remove-background,search-replace,search-recolor,upscale-fast,upscale-conservative,upscale-creative,control-sketch,control-structure,style-guide,style-transfer.- The tool takes a local PNG, JPEG or WebP (
image_path, plusmask_pathorstyle_image_pathwhere the service uses one) and saves the result as a PNG, the same waybedrock_generate_imagedoes. - Each service's required and optional fields live in
src/data/image-edit-models.json. A missing or unsupported field is refused before the billed call, and so is a file that is not one of those image formats. - Every operation was live-probed on 2026-11-01.
- The three upscalers return 4x images whose PNG is 22 to 43 MB, over Bedrock's 16 MB response limit, so they default to WebP (about 2.5 to 5.5 MB).
output_formatoverrides this per call.
- The tool takes a local PNG, JPEG or WebP (
- Gemma 4 31B (
gemma), Gemma 4 26B A4B (gemma-4-26b) and Gemma 4 E2B (gemma-4-e2b), served only on Bedrock's OpenAI-compatiblebedrock-mantleendpoint.- A new
"api": "mantle"registry field routes them tosrc/bedrock/mantle.ts, which calls Chat Completions athttps://bedrock-mantle.<region>.api.aws/openai/v1. The request is built by the samebuildConverseInput, so option checks, output caps and temperature rules match, and compare and judge mix them with Converse models. - Probed on 2026-11-01:
- plain calls,
temperature, system prompts and a JSON schema (response_format) all answered; effort(reasoning_effort) takeslowtoxhigh, whilemaxgets a 400 ("not supported with this model");- an abort at 400 ms settled at 411 ms.
- plain calls,
- These models need
AWS_BEARER_TOKEN_BEDROCK. AWS also accepts SigV4 on mantle, but this server does not sign mantle requests and says so when the token is missing. - Grok 4.3, also mantle-only, is left out because Grok 4.7 supersedes it.
- A new
- Nine text models, each live-probed on 2026-11-01 for a plain call,
temperature,schemaandeffortbefore it was added:- GPT-6.1 Sol (
gpt-sol) and GPT-6.1 Terra (gpt-terra), Grok 4.7 (grok) and Kimi K3 (kimi). All four takeschemaandeffortand rejecttemperature("This model doesn't support the temperature field"), so they are flaggednoTemperature. - MiniMax M2.5 (
minimax), Nemotron 3 Super 120B (nemotron), Ministral 3 14B (ministral) and Qwen3 Next 80B A3B (qwen-next). These takeschemaandtemperature, noteffort. - Magistral Small (
magistral). It is not flaggedstructured_output: it acceptedoutputConfig.textFormatbut ignored it, answering with fenced JSON and a different key.
- GPT-6.1 Sol (
- Cohere Embed v4 (
cohere.embed-v4:0, aliascohere, 1536 dimensions). Its response keys the vectors by type ({ float: [...] }) even when no type is requested, unlike v3's bare array;embed()reads both, and it also recognizes theus.cohere.inference profile. scripts/check-lifecycle.mjs(npm run lifecycle) fails when a registry model is listed on AWS's public model lifecycle page. It needs no credentials..github/workflows/lifecycle.ymlruns it weekly and on PRs that touchsrc/data/. Against the 0.6.0 registry it flags exactly the five Legacy models below.- Cancellation: every tool except
bedrock_generate_videopasses the MCP request's abort signal to Bedrock, so a cancelled or timed-out tool call stops the request. Measured on 2026-11-01: a Nova Micro call aborted at 400 ms settled at 412 ms; the same call left to finish took 6.4 s. Video starts are excluded on purpose, because cancelling a sentStartAsyncInvokecould leave a billed job with no ARN to track.
Removed
- Breaking: Llama 4 Maverick (
llama4), Llama 4 Scout, Llama 3.3 70B, Devstral 2 and DeepSeek R1.- AWS moved all five to Legacy on 2026-09-30, with EOL on 2027-03-30, and "new customers can't use Legacy models". So the
llama4examples in the tool descriptions and README failed for new accounts. - AWS lists no Active Meta model, so Llama leaves the catalog.
- AWS moved all five to Legacy on 2026-09-30, with EOL on 2027-03-30, and "new customers can't use Legacy models". So the
- Breaking: these Active but superseded models are removed:
- Claude Opus 4.8 and 4.7 (
claude-opus-4.8,claude-opus-4.7); Opus 5 and 5.5 cover them, and Opus 4.6 stays as the Opus that still takestemperature. - Mistral Small 2402 (
mistral-small), now covered by Ministral 3. - Titan Text Embeddings V1 (
titan-v1). - Cohere Embed English and Multilingual v3. Their aliases (
cohere-en,cohere-english,cohere-multi,cohere-multilingual) now resolve to Embed v4.
- Claude Opus 4.8 and 4.7 (
- The raw-HTTP bearer-token fallback (
withBearerFallback,bedrockFetch) and its 7 tests.- With
AWS_BEARER_TOKEN_BEDROCKset, the SDK already sends that token, ahead of any IAM keys, so the retry resent the same rejected credential and could never succeed. - Its
bedrock_video_statusbranch also requested/async-invoke/<id>where the API takes the full ARN.
- With
- The Amazon image request shape and the image registry's
formatfield. Every Active text-to-image model on Bedrock is Stability's (Nova Canvas reached EOL on 2026-09-30), so an image id outside the registry now gets the Stability shape too. - The
rimrafdev dependency;npm run cleanusesfs.rmSync.
Fixed
- A set-but-empty
AWS_BEARER_TOKEN_BEDROCKbroke every call with "request could not be signed withtokensince thetokenis not defined".- This happens with an MCP config that passes
"${AWS_BEARER_TOKEN_BEDROCK}"from a client whose own environment lacks the variable: the SDK picks bearer auth whenever the variable exists. - The server now drops an empty value at startup, so the IAM credential chain answers.
- This happens with an MCP config that passes
bedrock_comparereported success when every model failed; it now setsisErrorin that case. Partial failures still return the answers that came back.- Two images finished in the same millisecond overwrote each other. File names now end in an 8-character random suffix.
max_tokensonbedrock_ask,bedrock_compareandbedrock_judgeaccepted fractions, andseedonbedrock_generate_imageaccepted fractions and negatives, all of which Bedrock then rejected. The tool schemas now refuse them.- Titan embeddings ran one request after another (50 texts, 50 round trips in a row). They now run 8 at a time.
server.jsondescribed 26 text models; the registry has 33 now.
Security
npm auditwent from 5 findings to 0, closing the 6 open Dependabot alerts:- body-parser 2.3.0 -> 2.4.0 (GHSA-27r3-jxmh-p828, high) and fast-uri 3.1.8 -> 3.1.9 (4 advisories), both through
@modelcontextprotocol/sdk1.32.2. - glob and minimatch left with rimraf.
- body-parser 2.3.0 -> 2.4.0 (GHSA-27r3-jxmh-p828, high) and fast-uri 3.1.8 -> 3.1.9 (4 advisories), both through
- Manifest floors raised to
@aws-sdk/client-bedrock-runtime^3.1161.0 and@modelcontextprotocol/sdk^1.32.2.
Changed
- One
loadRegistryandindexByAliasinsrc/models.tsreplace the four copies of registry loading and alias indexing. OneerrorResultinsrc/index.tsreplaces six copies of the error branch.firstImageandsaveImageinsrc/bedrock/image.tsare shared by image generation and editing. - Examples in the tool descriptions, README and scripts use
kimiinstead ofllama4. Theeffortandschemadescriptions list the models that take them. release.ymlskipsnpm publishwhen npm already has the version. The first version is published by hand (npm attaches a trusted publisher only to an existing package), so its tag now goes on to the MCP Registry step, and a rerun after a failed registry step no longer stops at npm.package.jsonrepository.urluses thegit+httpsform npm normalizes to, sonpm publishno longer auto-corrects it.