Easy Migration - Change Just One line code
Migrate seamlessly by simply updating your baseURL no code refactoring needed.
AGIHALOUnified model routing, memory, and spend controlsLLM Router
LLM Router
for AI Agents
Always-On Decentralized Agents and AI
SaaS with USDC Payments on X402 Rails
Search by task name
API Keys
HALO APIKEYS
API Usage$
Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeDaily Revenue$Call volumeCost
Monthly Revenue$Call volumeCost
API KEYS4**************
Easy Migration Guide with Minimal Code Changes
Easy Migration with Just a BaseURL Change and X402 Logic for Agents Unified in One SDK
Unified SDK for Agents
All X402 payment logic and agent capabilities consolidated in a single SDK.
Autonomous X402 Flow
Reduce manual tasks with AI‑driven workflows (like expense categorization or time tracking).
Monthly revenue
Real-time Usage Tracking
Monitor API call volume by type with detailed usage analytics and cost breakdown.
Easy Migration
Migrate instantly by changing just the baseURL—zero refactoring needed.
All-in-One SDK
Access complete agent capabilities and X402 payment rails through a single SDK.
Autonomous AGI Economy
Decentralized agents that pay as they go, managing their own LLM expenses seamlessly.
X402 Automation
Autonomous payment handling: request, pay, and retry seamlessly.
The Autonomous API Management Layer for AI Agents
Issue unlimited API keys to track and limit costs for every individual agent precisely.
Fund agents with USDC and
automate model spend
X402 lets agents fund AGIHALO service usage and retry supported routes automatically.
Nano Banana 2.1
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeGemini 3.8 Flash TTS
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeReview customer support tickets and identify recurring issues
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeUpdate pricing page layout
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeGemini 3.8 Flash
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeGemini 4 Argon
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeGemini 3.6 Flash
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeGemini 3.5 Flash-Lite
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeNano Banana 2 Lite
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeUpdate pricing page layout
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by TypeUpdate pricing page layout
$Gemini 3 pro
Gemini 2.5 Flash Image Preview
Calls by Typex402 Rail Integration
Enable real-time micro-payments based on your actual usage on the x402 rail for maximum cost efficiency.
Agent Halo System
Empower Agents to autonomously manage and refuel their own LLM credits through integrated sub-LLM systems.
Managed Model Routing
Use one AGIHALO service layer for supported models, spend controls, metering, and agent memory.
USDC Decentralization
Shift payment authority to Agents and DAOs using USDC, bypassing the need for traditional credit card dependencies.
Build A.i Agent with AgiHalo
Agent Friendly LLM ROUTER WITH HALO
integrations and growing
LLM MODELS
Sync your code, issues, and deployments directly into your workflow.
Comunity
Connect support and CRM tools to turn conversations into action.
LONG-TERM MEMORY
Give every agent a memory that lasts.
Capture conversations per end user, organize them into summaries and topics, and retrieve only the context needed for the next response.
- End-user scoped
Keep every customer memory isolated by project and user.
- Topic-aware recall
Return relevant summaries and linked conversation context.
- SDK-first capture
Add memory to Node or Python agents without changing providers.
AGIHALO MEMORYLONG-TERM MEMORY WORKSPACE
MEMORY PROJECTCustomer Support Agentsupport-agent-memory / project memory data management
Users
3user_013 topics / 24 raw
user_025 topics / 41 raw
user_032 topics / 17 raw
Summary
user_01Product builder evaluating agent infrastructure. Prefers concise, code-first answers with predictable implementation costs.
user user_01last raw now
Topics
3 classifiedanswer_preferences
Concise, code-first responses
12 rawActive
product_context
Autonomous support agent
8 rawActive
billing_history
Weekly cost limit
4 rawActive
Raw Data
24 entriesAnswer preferencesanswer_preferences / raw 34dc8a1fActive
Product requirementsproduct_context / raw 782cb019Active
Execution
recentcapturedone
09:42 / user_01retrievedone
09:44 / user_01HALO MEMBERSHIP
Memory + Auth pricing
One membership for long-term memory and distinct Auth users per UTC month.
Halo Dev
$0/ month
For prototypes with managed memory and authentication.
- 50,000 Auth MAU
- 1 project
- 100 end users
- 1,000 writes / month
- 500 retrieves / month
- 100 MB storage
- 30-day retention
Halo Starter
$28/ month
For production apps growing memory and signed-in users.
- 50,000 Auth MAU
- 3 projects
- 1,000 end users
- 25,000 writes / month
- 5,000 retrieves / month
- 1 GB storage
- 90-day retention
Halo ProPOPULAR
$79/ month
For multi-agent products with durable context and Auth.
- 50,000 Auth MAU
- 20 projects
- 20,000 end users
- 250,000 writes / month
- 50,000 retrieves / month
- 10 GB storage
- 365-day retention
Halo Business
$249/ month
For large-scale identity and memory workloads.
- 50,000 Auth MAU
- 100 projects
- 100,000 end users
- 1,000,000 writes / month
- 250,000 retrieves / month
- 50 GB storage
- Extended retention
UPSTREAM REFERENCE RATES
Gemini pricing reference
Published Google standard rates for comparison only. AGIHALO service credits are not Google or Gemini API credits.
Official pricing| Model | Input rate | Output rate | Search after free quota | Source |
|---|---|---|---|---|
| Gemini 4 Argongemini-4-argon | $2.00 → $4.00Rate changes Dec 1, 2026 UTC | $10.00 → $20.00Rate changes Dec 1, 2026 UTC | $14 / 1K queries | Google list |
| Gemini 3.8 Flashgemini-3.8-flash | $0.75 → $1.50Rate changes Jan 1, 2027 UTC | $3.75 → $7.50Rate changes Jan 1, 2027 UTC | $14 / 1K queries | Google list |
| Gemini 3.6 Flashgemini-3.6-flash | $0.75 → $1.50Rate changes Jan 1, 2027 UTC | $3.75 → $7.50Rate changes Jan 1, 2027 UTC | $14 / 1K queries | Google list |
| Gemini 3.5 Flash-Litegemini-3.5-flash-lite | $0.30Standard | $2.50Standard | $14 / 1K queries | Google list |
| Gemini 3.1 Pro Previewgemini-3.1-pro-preview | $2.00 / $4.00<=200K / >200K | $12.00 / $18.00<=200K / >200K | $14 / 1K queries | Google list |
| Gemini 3.1 Flash-Litegemini-3.1-flash-lite | $0.25Standard | $1.50Standard | $14 / 1K queries | Google list |
| Gemini 3 Flash Previewgemini-3-flash-preview | $0.50Legacy preview | $3.00Legacy preview | $14 / 1K queries | Google list |
| Nano Banana 2.1gemini-nano-banana-2.1 | $1.501K image ≈ $0.0336 | $7.50 text / $30 image tokens1K image ≈ $0.0336 | $14 / 1K queries | Google list |
| Nano Banana 2 Litegemini-3.1-flash-lite-image | $0.251K image ≈ $0.0336 | $1.50 text / $30 image tokens1K image ≈ $0.0336 | $14 / 1K queries | Google list |
| Gemini 3 Pro Imagegemini-3-pro-image | $2.001K-2K image output | $12 text / $0.134 image1K-2K image output | $14 / 1K queries | Google list |
| Gemini 3.8 Flash TTSgemini-3.8-flash-tts | $0.50 → $1.00Rate changes Jan 1, 2027 UTC | $9.00 → $18.00 audioRate changes Jan 1, 2027 UTC | Not available | Google list |
| Gemini 3.8 Flash-Lite TTSgemini-3.8-flash-lite-tts | $0.50 → $1.00Rate changes Jan 1, 2027 UTC | $6.00 → $12.00 audioRate changes Jan 1, 2027 UTC | Not available | Google list |
| Gemini 2.5 Progemini-2.5-pro | $1.25 / $2.50<=200K / >200K · existing projects only | $10.00 / $15.00<=200K / >200K · existing projects only | $35 / 1K grounded prompts | Google list |
| Gemini 2.5 Flashgemini-2.5-flash | $0.30Existing Google projects only | $2.50Existing Google projects only | $35 / 1K grounded prompts | Google list |
| Gemini 2.5 Flash Imagegemini-2.5-flash-image | $0.30Existing Google projects only | $0.039 / imageExisting Google projects only | Not available | Google list |
OPENAI · TEXT & REASONING
OpenAI GPT pricing reference
OpenAI Standard rates in USD per 1M tokens. Paired values show short and long context rates; long-context pricing applies above 272K input tokens.
Official pricing| Model | Input | Cached input | Cache write | Output | Rate tier |
|---|---|---|---|---|---|
| GPT-6 Astragpt-6-astra | $10.00 / $20.00per 1M tokens | $1.00 / $2.00per 1M tokens | $12.50 / $25.00per 1M tokens | $50.00 / $75.00per 1M tokens | Short / long context |
| GPT-6.1 Solgpt-6.1-sol | $2.00 / $4.00per 1M tokens | $0.10 / $0.20per 1M tokens | $2.50 / $5.00per 1M tokens | $10.00 / $15.00per 1M tokens | Short / long context |
| GPT-6.1 Terragpt-6.1-terra | $1.00 / $2.00per 1M tokens | $0.05 / $0.10per 1M tokens | $1.25 / $2.50per 1M tokens | $5.00 / $7.50per 1M tokens | Short / long context |
| GPT-6 Lunagpt-6-luna | $0.10 / $0.20per 1M tokens | $0.01 / $0.02per 1M tokens | $0.125 / $0.25per 1M tokens | $0.50 / $0.75per 1M tokens | Short / long context |
| GPT-6 Solgpt-6-sol | $2.00 / $4.00per 1M tokens | $0.20 / $0.40per 1M tokens | $2.50 / $5.00per 1M tokens | $10.00 / $15.00per 1M tokens | Previous Sol · short / long context |
| GPT-5.6 Solgpt-5.6 → gpt-5.6-sol | $4.00 / $8.00per 1M tokens | $0.40 / $0.80per 1M tokens | $5.00 / $10.00per 1M tokens | $20.00 / $30.00per 1M tokens | Promotional short / long context |
| GPT-5.6 Terragpt-5.6-terra | $2.00 / $4.00per 1M tokens | $0.20 / $0.40per 1M tokens | $2.50 / $5.00per 1M tokens | $12.00 / $18.00per 1M tokens | Short / long context |
| GPT-5.6 Lunagpt-5.6-luna | $0.20 / $0.40per 1M tokens | $0.02 / $0.04per 1M tokens | $0.25 / $0.50per 1M tokens | $1.20 / $1.80per 1M tokens | Short / long context |
| GPT-5.5gpt-5.5 | $5.00 / $10.00per 1M tokens | $0.50 / $1.00per 1M tokens | —per 1M tokens | $30.00 / $45.00per 1M tokens | Short / long context |
| GPT-5.5 Progpt-5.5-pro | $30.00 / $60.00per 1M tokens | —per 1M tokens | —per 1M tokens | $180.00 / $270.00per 1M tokens | Responses API only |
| GPT-5.4gpt-5.4 | $2.50 / $5.00per 1M tokens | $0.25 / $0.50per 1M tokens | —per 1M tokens | $15.00 / $22.50per 1M tokens | Short / long context |
| GPT-5.4 Minigpt-5.4-mini | $0.75per 1M tokens | $0.075per 1M tokens | —per 1M tokens | $4.50per 1M tokens | Standard |
| GPT-5.4 Nanogpt-5.4-nano | $0.20per 1M tokens | $0.02per 1M tokens | —per 1M tokens | $1.25per 1M tokens | Standard |
| GPT-5.4 Progpt-5.4-pro | $30.00 / $60.00per 1M tokens | —per 1M tokens | —per 1M tokens | $180.00 / $270.00per 1M tokens | Responses API only |
| GPT-5.2gpt-5.2 | $1.75per 1M tokens | $0.175per 1M tokens | —per 1M tokens | $14.00per 1M tokens | Standard |
| GPT-5.2 Progpt-5.2-pro | $21.00per 1M tokens | —per 1M tokens | —per 1M tokens | $168.00per 1M tokens | Responses API only |
| GPT-5.1gpt-5.1 | $1.25per 1M tokens | $0.125per 1M tokens | —per 1M tokens | $10.00per 1M tokens | Standard |
OPENAI · IMAGE GENERATION
OpenAI image pricing reference
Standard rates in USD per 1M text or image tokens. Actual per-image cost varies by size and quality.
Official pricing| Model | Text tokens | Image tokens | Tier |
|---|---|---|---|
| GPT Image 2.5 Sunburstgpt-image-2.5-sunburst | $5.00 input$1.25 cached (Responses only) | $8.00 input · $30.00 output$2.00 cached (Responses only) | Latest · editing precision |
| GPT Image 2.5 Flaregpt-image-2.5-flare | $5.00 input$1.25 cached (Responses only) | $8.00 input · $30.00 output$2.00 cached (Responses only) | Latest · fast generation |
| GPT Image 2gpt-image-2 | $5.00 input$1.25 cached | $8.00 input · $30.00 output$2.00 cached | Previous |
| GPT Image 1.5gpt-image-1.5 | $5.00 input · $10.00 output$1.25 cached | $8.00 input · $32.00 output$2.00 cached | Standard |
| GPT Image 1 Minigpt-image-1-mini | $2.00 input$0.20 cached | $2.50 input · $8.00 output$0.25 cached | Economy |
| GPT Image 1gpt-image-1 | $5.00 input$1.25 cached | $10.00 input · $40.00 output$2.50 cached | Previous |
| ChatGPT Image Latestchatgpt-image-latest | $5.00 input · $10.00 output$1.25 cached | $8.00 input · $32.00 output$2.00 cached | ChatGPT alias |
UPSTREAM REFERENCE RATES
Anthropic Claude pricing reference
USD per 1M tokens. Haiku 5.5 paired rates apply up to / above 100K input tokens.
Official pricing| Model | Input rate | Cached input | Output rate | Rate tier |
|---|---|---|---|---|
| Claude Fable 5.1claude-fable-5-1 | $10.00 | $0.25 | $50.00 | Standard |
| Claude Opus 5.5claude-opus-5-5 | $4.00 | $0.20 | $20.00 | Standard |
| Claude Sonnet 5.5claude-sonnet-5-5 | $2.00 | $0.10 | $10.00 | Standard |
| Claude Haiku 5.5claude-haiku-5-5 | $0.10 / $0.50 | $0.01 / $0.05 | $0.50 / $2.50 | Up to / above 100K |
UPSTREAM REFERENCE RATES
DeepSeek pricing reference
USD per 1M tokens. Paired rates show peak / off-peak; off-peak rates are 50% lower.
Official pricing| Model | Input rate | Cached input | Output rate | Rate tier |
|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017deepseek-flash | $0.30 / $0.15 | $0.006 / $0.003 | $1.20 / $0.60 | Text and image input |
| DeepSeek V4 Prodeepseek-v4-pro | $1.32 / $0.66 | $0.044 / $0.022 | $3.96 / $1.98 | Text input |
UPSTREAM REFERENCE RATES
Fish Audio pricing reference
USD per 1M UTF-8 input bytes. The input rate includes generated speech.
Official pricing| Model | Input rate | Cached input | Output rate | Rate tier |
|---|---|---|---|---|
| Drama 3drama-3 | $10.00 → $50.00 | — | Included | Rate changes Nov 12, 2026 at 17:00 UTC |
| S2.1 Pros2.1-pro | $15.00 | — | Included | Standard |
| S2 Pros2-pro | $15.00 | — | Included | Previous |
LLM POWER PROVIDERS
Provide open-source inference capacity.
Publish an OpenAI-compatible inference endpoint backed by your own compute. Halo tests the connection, checks the advertised models, and lets you prepare model-specific marketplace pricing.
Start providingCustom-provider registration tests every advertised Halo model with a real minimal request and shows its response time. Buyer requests reserve their maximum cost before dispatch, then settle with Halo text units calculated from the canonical request and visible assistant text, without switching to another seller. Individual KYC is required, and the default split is a 10% Halo fee and 90% provider earnings before upstream costs.Hosted API keys — not yet live
Hosted seller credentials may appear during rollout, but they are not currently connected to buyer routing or seller earnings. Do not register one expecting traffic or revenue.
Provide compute capacity
Publish an OpenAI-compatible chat endpoint backed by your GPU or inference server, then enable the models you can serve.
- 01Test & publish custom capacityVerify the custom endpoint and publish enabled model listings.
- 02Serve routed requestsKeep each custom model healthy while live buyer traffic is routed to eligible custom listings.
- 03Track earningsSuccessful routed usage is metered in Halo text units and reflected in provider earnings.
Frequently asked questions
Get answers to commonly asked questions.
Is this different from existing Gemini features?
AGIHALO keeps the underlying model capabilities while adding unified routing, API spend governance, long-term memory, and X402 payment automation for agents.
Is payment only possible through cryptocurrency?
No. Account credits can be purchased by card or with USDC. X402 enables autonomous agent payments over supported routes.
Are AGIHALO credits Google or Gemini API credits?
No. AGIHALO credits pay only for AGIHALO routing, memory, metering, and related infrastructure. They cannot be transferred to or redeemed with Google or another model vendor.
How is my input data used?
Requests are processed to provide model routing, usage metering, and billing. Memory content is stored only when you enable long-term memory, as described in our Privacy Policy.
What are the main advantages of the Halo system?
AGIHALO gives agents one API for model access, per-key cost controls, durable end-user memory, provider monetization, and automated X402 payment flows.
LLM Router for AI Agents
Routing, long-term memory, and autonomous USDC payments for always-on agents
Get started