tokenbee-sdk 1.3.0__tar.gz → 2.0.1__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (11) hide show
  1. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/PKG-INFO +10 -10
  2. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/README.md +9 -9
  3. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/pyproject.toml +1 -1
  4. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/setup.py +1 -1
  5. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee/__init__.py +35 -18
  6. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/PKG-INFO +10 -10
  7. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/setup.cfg +0 -0
  8. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/SOURCES.txt +0 -0
  9. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/dependency_links.txt +0 -0
  10. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/requires.txt +0 -0
  11. {tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/top_level.txt +0 -0
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: tokenbee-sdk
3
- Version: 1.3.0
3
+ Version: 2.0.1
4
4
  Summary: Official Python SDK for TokenBee LLM inference gateway and observability.
5
5
  Author: TokenBee Inc.
6
6
  Author-email: "TokenBee Inc." <founders@tokenbee.io>
@@ -18,13 +18,13 @@ Dynamic: requires-python
18
18
 
19
19
  # TokenBee Python SDK
20
20
 
21
- Official Python SDK for [TokenBee](https://tokenbee.io) - The Intelligent LLM Inference Gateway with Observability, Compression, and Privacy.
21
+ Official Python SDK for [TokenBee](https://tokenbee.io) — AI interaction capture, audit, replay, and optimization.
22
22
 
23
23
  ## Features
24
24
 
25
25
  - **Unified API**: Access multiple LLM providers (OpenAI, Anthropic, Google, Mistral, etc.) through a single interface.
26
26
  - **Intelligent Compression**: Reduce token usage and latency with context-aware compression.
27
- - **Privacy Guard**: Automatic PII masking and privacy-preserving inference.
27
+ - **Configurable Capture**: Retain interaction content when you need it; turn capture off per request or globally.
28
28
  - **Built-in Observability**: Automatic tracking of latency, costs, and token usage.
29
29
 
30
30
  ## Links
@@ -54,7 +54,7 @@ client = TokenBee(
54
54
 
55
55
  # Send a request
56
56
  response = client.send(
57
- model=TokenBeeModel.OPENAI_GPT_4O,
57
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
58
58
  input={
59
59
  "messages": [
60
60
  {"role": "user", "content": "Explain quantum entanglement in simple terms."}
@@ -77,12 +77,12 @@ You can specify the compression rate and method per request. TokenBee uses an in
77
77
 
78
78
  ```python
79
79
  response = client.send(
80
- model=TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET,
80
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
81
81
  input={
82
82
  "messages": [...],
83
83
  "compression": "auto", # "auto" (default), "on", or "off"
84
84
  "rate": CompressionRate.HIGH, # MEDIUM (0.5), HIGH (0.33), etc.
85
- "privacy": True
85
+ "capture": False # per-request: do not retain message/response bodies
86
86
  }
87
87
  )
88
88
  ```
@@ -91,15 +91,15 @@ response = client.send(
91
91
  - **`rate`**: Controls the aggressiveness of compression. `HIGH` aims for ~67% token reduction.
92
92
  - **`sessionId`**: (Optional) String ID to group multiple requests into a single replayable session in the dashboard.
93
93
  - **`userId`**: (Optional) String ID to track usage and costs per unique end-user.
94
- - **`privacy`**: (Optional) Set to `True` to disable payload logging and session replays for this request. Metadata (latency, tokens) will still be recorded for observability.
94
+ - **`capture`**: (Optional) Per-request override. `False` does not retain messages/responses (metadata like tokens/cost/latency can still be recorded). Precedence: per-request → project setting → default on.
95
95
 
96
96
  ### Supported Models
97
97
 
98
98
  The SDK provides a `TokenBeeModel` enum with popular models:
99
99
 
100
- - `TokenBeeModel.OPENAI_GPT_4O`
101
- - `TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET`
102
- - `TokenBeeModel.GEMINI_2_0_FLASH`
100
+ - `TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4`
101
+ - `TokenBeeModel.OPENAI_GPT_5_MINI`
102
+ - `TokenBeeModel.GROQ_GPT_OSS_20B`
103
103
  - ... and many others.
104
104
 
105
105
  ## License
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/README.md RENAMED
@@ -1,12 +1,12 @@
1
1
  # TokenBee Python SDK
2
2
 
3
- Official Python SDK for [TokenBee](https://tokenbee.io) - The Intelligent LLM Inference Gateway with Observability, Compression, and Privacy.
3
+ Official Python SDK for [TokenBee](https://tokenbee.io) — AI interaction capture, audit, replay, and optimization.
4
4
 
5
5
  ## Features
6
6
 
7
7
  - **Unified API**: Access multiple LLM providers (OpenAI, Anthropic, Google, Mistral, etc.) through a single interface.
8
8
  - **Intelligent Compression**: Reduce token usage and latency with context-aware compression.
9
- - **Privacy Guard**: Automatic PII masking and privacy-preserving inference.
9
+ - **Configurable Capture**: Retain interaction content when you need it; turn capture off per request or globally.
10
10
  - **Built-in Observability**: Automatic tracking of latency, costs, and token usage.
11
11
 
12
12
  ## Links
@@ -36,7 +36,7 @@ client = TokenBee(
36
36
 
37
37
  # Send a request
38
38
  response = client.send(
39
- model=TokenBeeModel.OPENAI_GPT_4O,
39
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
40
40
  input={
41
41
  "messages": [
42
42
  {"role": "user", "content": "Explain quantum entanglement in simple terms."}
@@ -59,12 +59,12 @@ You can specify the compression rate and method per request. TokenBee uses an in
59
59
 
60
60
  ```python
61
61
  response = client.send(
62
- model=TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET,
62
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
63
63
  input={
64
64
  "messages": [...],
65
65
  "compression": "auto", # "auto" (default), "on", or "off"
66
66
  "rate": CompressionRate.HIGH, # MEDIUM (0.5), HIGH (0.33), etc.
67
- "privacy": True
67
+ "capture": False # per-request: do not retain message/response bodies
68
68
  }
69
69
  )
70
70
  ```
@@ -73,15 +73,15 @@ response = client.send(
73
73
  - **`rate`**: Controls the aggressiveness of compression. `HIGH` aims for ~67% token reduction.
74
74
  - **`sessionId`**: (Optional) String ID to group multiple requests into a single replayable session in the dashboard.
75
75
  - **`userId`**: (Optional) String ID to track usage and costs per unique end-user.
76
- - **`privacy`**: (Optional) Set to `True` to disable payload logging and session replays for this request. Metadata (latency, tokens) will still be recorded for observability.
76
+ - **`capture`**: (Optional) Per-request override. `False` does not retain messages/responses (metadata like tokens/cost/latency can still be recorded). Precedence: per-request → project setting → default on.
77
77
 
78
78
  ### Supported Models
79
79
 
80
80
  The SDK provides a `TokenBeeModel` enum with popular models:
81
81
 
82
- - `TokenBeeModel.OPENAI_GPT_4O`
83
- - `TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET`
84
- - `TokenBeeModel.GEMINI_2_0_FLASH`
82
+ - `TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4`
83
+ - `TokenBeeModel.OPENAI_GPT_5_MINI`
84
+ - `TokenBeeModel.GROQ_GPT_OSS_20B`
85
85
  - ... and many others.
86
86
 
87
87
  ## License
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/pyproject.toml RENAMED
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
4
4
 
5
5
  [project]
6
6
  name = "tokenbee-sdk"
7
- version = "1.3.0"
7
+ version = "2.0.1"
8
8
  authors = [
9
9
  { name="TokenBee Inc.", email="founders@tokenbee.io" },
10
10
  ]
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/setup.py RENAMED
@@ -6,7 +6,7 @@ long_description = (this_directory / "README.md").read_text()
6
6
 
7
7
  setup(
8
8
  name="tokenbee-sdk",
9
- version="1.3.0",
9
+ version="2.0.1",
10
10
  packages=find_packages(),
11
11
  install_requires=["httpx>=0.23.0"],
12
12
  author="TokenBee Inc.",
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee/__init__.py RENAMED
@@ -1,6 +1,7 @@
1
1
  import os
2
2
  import httpx
3
3
  from enum import Enum
4
+ from typing import Optional
4
5
 
5
6
  class CompressionRate(str, Enum):
6
7
  LOW = "0.75"
@@ -20,28 +21,39 @@ class TokenBeeContext(str, Enum):
20
21
  CODE = "code"
21
22
 
22
23
  class TokenBeeModel(str, Enum):
23
- # OpenAI
24
- OPENAI_GPT_4_5 = "openai/gpt-4.5-preview"
24
+ # OpenAI — current GPT-5.6 / GPT-5 family
25
+ OPENAI_GPT_5_6_SOL = "openai/gpt-5.6-sol"
26
+ OPENAI_GPT_5_6_TERRA = "openai/gpt-5.6-terra"
27
+ OPENAI_GPT_5_6_LUNA = "openai/gpt-5.6-luna"
28
+ OPENAI_GPT_5 = "openai/gpt-5"
29
+ OPENAI_GPT_5_MINI = "openai/gpt-5-mini"
30
+ OPENAI_GPT_5_NANO = "openai/gpt-5-nano"
31
+ OPENAI_GPT_5_4 = "openai/gpt-5.4"
32
+ OPENAI_GPT_5_4_MINI = "openai/gpt-5.4-mini"
33
+ OPENAI_GPT_5_4_NANO = "openai/gpt-5.4-nano"
34
+ OPENAI_GPT_4_1 = "openai/gpt-4.1"
35
+ OPENAI_GPT_4_1_MINI = "openai/gpt-4.1-mini"
36
+ OPENAI_GPT_4_1_NANO = "openai/gpt-4.1-nano"
25
37
  OPENAI_GPT_4O = "openai/gpt-4o"
26
38
  OPENAI_GPT_4O_MINI = "openai/gpt-4o-mini"
39
+ OPENAI_O3 = "openai/o3"
40
+ OPENAI_O3_MINI = "openai/o3-mini"
41
+ OPENAI_O4_MINI = "openai/o4-mini"
27
42
  OPENAI_O1 = "openai/o1"
28
43
  OPENAI_O1_MINI = "openai/o1-mini"
29
- OPENAI_O3_MINI = "openai/o3-mini"
30
44
 
31
45
  # Anthropic
46
+ ANTHROPIC_CLAUDE_SONNET_4 = "anthropic/claude-sonnet-4-latest"
47
+ ANTHROPIC_CLAUDE_OPUS_4 = "anthropic/claude-opus-4-latest"
48
+ ANTHROPIC_CLAUDE_HAIKU_4 = "anthropic/claude-haiku-4-latest"
32
49
  ANTHROPIC_CLAUDE_3_7_SONNET = "anthropic/claude-3-7-sonnet-latest"
33
50
  ANTHROPIC_CLAUDE_3_5_SONNET = "anthropic/claude-3-5-sonnet-latest"
34
51
  ANTHROPIC_CLAUDE_3_5_HAIKU = "anthropic/claude-3-5-haiku-latest"
35
- ANTHROPIC_CLAUDE_3_OPUS = "anthropic/claude-3-opus-latest"
36
52
 
37
53
  # Google
38
- GEMINI_3_1_PRO = "google/gemini-3.1-pro"
39
- GEMINI_3_1_FLASH = "google/gemini-3.1-flash"
40
54
  GEMINI_2_5_PRO = "google/gemini-2.5-pro"
55
+ GEMINI_2_5_FLASH = "google/gemini-2.5-flash"
41
56
  GEMINI_2_0_FLASH = "google/gemini-2.0-flash"
42
- GEMINI_2_0_PRO = "google/gemini-2.0-pro-exp"
43
- GEMINI_1_5_PRO = "google/gemini-1.5-pro"
44
-
45
57
 
46
58
  # Mistral
47
59
  MISTRAL_LARGE = "mistral/mistral-large-latest"
@@ -54,10 +66,14 @@ class TokenBeeModel(str, Enum):
54
66
  PERPLEXITY_SONAR_PRO = "perplexity/sonar-pro"
55
67
  PERPLEXITY_SONAR_REASONING = "perplexity/sonar-reasoning"
56
68
 
57
- # Groq
58
- GROQ_LLAMA_3_3_70B = "groq/llama-3.3-70b-versatile"
59
- GROQ_LLAMA_3_1_8B = "groq/llama-3.1-8b-instant"
60
- GROQ_DEEPSEEK_R1_DISTILL = "groq/deepseek-r1-distill-llama-70b"
69
+ # Groq (current — Aug 2026)
70
+ # Note: model IDs after "groq/" are sent to Groq as-is (may include org prefix).
71
+ GROQ_GPT_OSS_120B = "groq/openai/gpt-oss-120b"
72
+ GROQ_GPT_OSS_20B = "groq/openai/gpt-oss-20b"
73
+ GROQ_GPT_OSS_SAFEGUARD_20B = "groq/openai/gpt-oss-safeguard-20b"
74
+ GROQ_QWEN3_6_27B = "groq/qwen/qwen3.6-27b"
75
+ GROQ_COMPOUND = "groq/groq/compound"
76
+ GROQ_COMPOUND_MINI = "groq/groq/compound-mini"
61
77
 
62
78
  # xAI
63
79
  XAI_GROK_3 = "xai/grok-3"
@@ -75,7 +91,7 @@ class TokenBee:
75
91
  context: str = TokenBeeContext.AUTO,
76
92
  model: str = "",
77
93
  provider: str = "",
78
- privacy: bool = False
94
+ capture: Optional[bool] = None
79
95
  ):
80
96
  self.api_key = api_key
81
97
  self.llm_key = llm_key
@@ -89,8 +105,9 @@ class TokenBee:
89
105
  "X-TokenBee-Rate": rate,
90
106
  "X-TokenBee-Strategy": strategy,
91
107
  "X-TokenBee-Context": context,
92
- "X-TokenBee-Privacy": str(privacy).lower()
93
108
  }
109
+ if capture is not None:
110
+ self.headers["X-TokenBee-Capture"] = str(capture).lower()
94
111
  if model:
95
112
  self.headers["X-TokenBee-Model"] = model
96
113
  if provider:
@@ -115,8 +132,8 @@ class TokenBee:
115
132
  headers["X-TokenBee-Strategy"] = str(input["strategy"])
116
133
  if "context" in input:
117
134
  headers["X-TokenBee-Context"] = str(input["context"])
118
- if "privacy" in input:
119
- headers["X-TokenBee-Privacy"] = str(input["privacy"]).lower()
135
+ if "capture" in input:
136
+ headers["X-TokenBee-Capture"] = str(input["capture"]).lower()
120
137
  if "sessionId" in input:
121
138
  headers["X-TB-Session-Id"] = str(input["sessionId"])
122
139
  if "userId" in input:
@@ -128,7 +145,7 @@ class TokenBee:
128
145
  payload.pop("rate", None)
129
146
  payload.pop("strategy", None)
130
147
  payload.pop("context", None)
131
- payload.pop("privacy", None)
148
+ payload.pop("capture", None)
132
149
  payload.pop("sessionId", None)
133
150
  payload.pop("userId", None)
134
151
 
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/tokenbee_sdk.egg-info/PKG-INFO RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: tokenbee-sdk
3
- Version: 1.3.0
3
+ Version: 2.0.1
4
4
  Summary: Official Python SDK for TokenBee LLM inference gateway and observability.
5
5
  Author: TokenBee Inc.
6
6
  Author-email: "TokenBee Inc." <founders@tokenbee.io>
@@ -18,13 +18,13 @@ Dynamic: requires-python
18
18
 
19
19
  # TokenBee Python SDK
20
20
 
21
- Official Python SDK for [TokenBee](https://tokenbee.io) - The Intelligent LLM Inference Gateway with Observability, Compression, and Privacy.
21
+ Official Python SDK for [TokenBee](https://tokenbee.io) — AI interaction capture, audit, replay, and optimization.
22
22
 
23
23
  ## Features
24
24
 
25
25
  - **Unified API**: Access multiple LLM providers (OpenAI, Anthropic, Google, Mistral, etc.) through a single interface.
26
26
  - **Intelligent Compression**: Reduce token usage and latency with context-aware compression.
27
- - **Privacy Guard**: Automatic PII masking and privacy-preserving inference.
27
+ - **Configurable Capture**: Retain interaction content when you need it; turn capture off per request or globally.
28
28
  - **Built-in Observability**: Automatic tracking of latency, costs, and token usage.
29
29
 
30
30
  ## Links
@@ -54,7 +54,7 @@ client = TokenBee(
54
54
 
55
55
  # Send a request
56
56
  response = client.send(
57
- model=TokenBeeModel.OPENAI_GPT_4O,
57
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
58
58
  input={
59
59
  "messages": [
60
60
  {"role": "user", "content": "Explain quantum entanglement in simple terms."}
@@ -77,12 +77,12 @@ You can specify the compression rate and method per request. TokenBee uses an in
77
77
 
78
78
  ```python
79
79
  response = client.send(
80
- model=TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET,
80
+ model=TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4,
81
81
  input={
82
82
  "messages": [...],
83
83
  "compression": "auto", # "auto" (default), "on", or "off"
84
84
  "rate": CompressionRate.HIGH, # MEDIUM (0.5), HIGH (0.33), etc.
85
- "privacy": True
85
+ "capture": False # per-request: do not retain message/response bodies
86
86
  }
87
87
  )
88
88
  ```
@@ -91,15 +91,15 @@ response = client.send(
91
91
  - **`rate`**: Controls the aggressiveness of compression. `HIGH` aims for ~67% token reduction.
92
92
  - **`sessionId`**: (Optional) String ID to group multiple requests into a single replayable session in the dashboard.
93
93
  - **`userId`**: (Optional) String ID to track usage and costs per unique end-user.
94
- - **`privacy`**: (Optional) Set to `True` to disable payload logging and session replays for this request. Metadata (latency, tokens) will still be recorded for observability.
94
+ - **`capture`**: (Optional) Per-request override. `False` does not retain messages/responses (metadata like tokens/cost/latency can still be recorded). Precedence: per-request → project setting → default on.
95
95
 
96
96
  ### Supported Models
97
97
 
98
98
  The SDK provides a `TokenBeeModel` enum with popular models:
99
99
 
100
- - `TokenBeeModel.OPENAI_GPT_4O`
101
- - `TokenBeeModel.ANTHROPIC_CLAUDE_3_5_SONNET`
102
- - `TokenBeeModel.GEMINI_2_0_FLASH`
100
+ - `TokenBeeModel.ANTHROPIC_CLAUDE_SONNET_4`
101
+ - `TokenBeeModel.OPENAI_GPT_5_MINI`
102
+ - `TokenBeeModel.GROQ_GPT_OSS_20B`
103
103
  - ... and many others.
104
104
 
105
105
  ## License
{tokenbee_sdk-1.3.0 → tokenbee_sdk-2.0.1}/setup.cfg RENAMED
File without changes