0.154.1 (2026-07-29)

🐛 Bug Fixes

• web: refresh session activity via SessionWrapper to prevent premature cleanup (#969) (45bf156), closes #967

♻️ Code Refactoring

• clean up comments and formatting in anthropic_messages.go (46f3247)
• remove dead Config getters and unused ToolCallRenderer render chain (#971) (4b3b76e), closes #968

📚 Documentation

• examples: key-free chat and tracing demo for the k8s example (#970) (9dbc5eb)

🔧 Build System

• deps: bump docker/login-action from 4.4.0 to 4.5.2 in the github-actions group (#965) (f75c346)
• deps: bump the gomod group with 2 updates (#964) (3e8e4e0)

👷 CI/CD

• claude: centralize claude.yml via reusable workflow (#966) (0798171)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.154.1 --help
npx @inference-gateway/cli@0.154.1 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.154.1
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.154.1

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.154.1

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.154.1"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.154.1

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.154.0 (2026-07-28)

🚀 Features

• chat: add /effort command for runtime anthropic reasoning effort (#963) (f7a9ef7), references #962 inference-gateway/schemas#161 inference-gateway/sdk#148
• route anthropic models through gateway /v1/messages via removable adapter (#962) (6c99269), closes #942, references #941

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.154.0 --help
npx @inference-gateway/cli@0.154.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.154.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.154.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.154.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.154.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.154.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.153.3 (2026-07-27)

👷 CI/CD

• claude: centralize claude.yml via reusable workflow (#960) (6891637)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.153.3 --help
npx @inference-gateway/cli@0.153.3 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.153.3
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.153.3

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.153.3

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.153.3"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.153.3

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.153.2 (2026-07-27)

🐛 Bug Fixes

• ci: update maintainer app ID to client ID in workflows and documentation (c09f400)

🧹 Maintenance

• deps: bump infer CLI v0.151.0 -> v0.153.1 (#958) (3c7351e)
• deps: bump sdk v1.26.0 -> v1.28.0 and adk v0.23.4 -> v0.25.0 (#959) (fc4305c)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.153.2 --help
npx @inference-gateway/cli@0.153.2 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.153.2
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.153.2

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.153.2

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.153.2"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.153.2

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.153.1 (2026-07-26)

🐛 Bug Fixes

• make host.docker.internal resolvable for agent containers (Linux CI) (#957) (120d9d3)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.153.1 --help
npx @inference-gateway/cli@0.153.1 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.153.1
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.153.1

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.153.1

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.153.1"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.153.1

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.153.0 (2026-07-26)

🚀 Features

• bash: allow mkdir and ln in the read-only baseline (#956) (fdb78e8)

👷 CI/CD

• claude: centralize claude.yml via reusable workflow (#954) (df49f62)
• claude: centralize claude.yml via reusable workflow (#955) (f375507)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.153.0 --help
npx @inference-gateway/cli@0.153.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.153.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.153.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.153.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.153.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.153.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.152.0 (2026-07-22)

🚀 Features

• telemetry: include tool call ID in failed-tool log records (#953) (07ce6fa), closes #947
• examples: k3d Kubernetes example with operator CRDs, channels-manager, A2A agent, and telemetry (#944) (e12bff6), closes #943

🐛 Bug Fixes

• agent: discount cached prompt tokens in session cost accumulator (#951) (59e5496)

✅ Tests

• memory: fix flaky TestGitBackend TempDir cleanup (#952) (94d23fa), closes #950

🔧 Build System

• deps: bump google.golang.org/grpc from 1.81.1 to 1.82.1 (#945) (d08f94a)
• deps: bump the github-actions group with 2 updates (#949) (6ee94e0)
• deps: bump the gomod group with 3 updates (#948) (a6538a6)

👷 CI/CD

• deps: enable k3d in GitHub Actions workflow (3d9475b)

🧹 Maintenance

• deps: add k3d package to environment (5f9c44d)
• flox: update infer to v0.151.0 (4acb4ee)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.152.0 --help
npx @inference-gateway/cli@0.152.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.152.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.152.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.152.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.152.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.152.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.151.0 (2026-07-21)

🚀 Features

• cache-stable system prompt, gateway-sourced model metadata, and cached-token visibility (#941) (ef0f69e), references #938 #937 #207

🧹 Maintenance

• pricing: add kimi-k3 model support (3038c65)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.151.0 --help
npx @inference-gateway/cli@0.151.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.151.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.151.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.151.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.151.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.151.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.150.0 (2026-07-19)

🚀 Features

• gpu: on-demand GPU provisioning with llama.cpp via RunPod (#940) (87bb72a), references #939 #939 inference-gateway/cli#939
• agent: reasoning_effort config + working-offline example (#937) (cadd230)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.150.0 --help
npx @inference-gateway/cli@0.150.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.150.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.150.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.150.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.150.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.150.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)

0.149.0 (2026-07-18)

🚀 Features

• telemetry: add gen_ai.tool.call.id attribute to tool execution spans (#931) (b695e16), closes #930
• providers: add llamacpp provider support (#929) (b9905ff), closes #928
• a2a: add recurring liveness probes for A2A agents (#936) (01ff0d7), closes #932
• channels: rich Telegram messaging - photos, markdown, slash commands, and per-conversation /stats (#927) (e57a10c)
• a2a: stream agent image pull progress and color-code readiness indicator (#933) (882e8c2), references #932

♻️ Code Refactoring

• agents: make model requirement metadata-driven for known agents (#935) (e0b8df8), closes #934

🧹 Maintenance

• telegram: add INFER_SPEECH_TO_TEXT_MODEL configuration (fadde8f)

📦 Installation

npm / npx (Recommended)

Most developers already have Node.js - run infer without installing anything. npx downloads the matching native binary on first use:

npx @inference-gateway/cli@0.149.0 --help
npx @inference-gateway/cli@0.149.0 chat

Or install it globally:

npm install -g @inference-gateway/cli@0.149.0
infer --help

Not recommended for production - prefer the install script, container image, or Nix flake below.

Quick Install (Install Script)

Install the latest version using our install script:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash

Or install a specific version:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --version v0.149.0

Custom installation directory:

curl -fsSL https://raw.githubusercontent.com/inference-gateway/cli/main/install.sh | bash -s -- --install-dir $HOME/.local/bin

Nix Flake

Run directly without installing:

nix run github:inference-gateway/cli/v0.149.0

Or pin it in a Flox manifest (.flox/env/manifest.toml):

[install]
infer.flake = "github:inference-gateway/cli/v0.149.0"

Container Image

docker run --rm -it ghcr.io/inference-gateway/cli:0.149.0

Binary Download

Download the appropriate binary for your platform from the release assets, or see the verification guide.

This release is also available on:

• npm package (@latest dist-tag)