Here’s a conversation that happens in every company evaluating AI tools:
Team lead: “We want to use JieGou for our marketing workflows.” Finance: “We already pay $2,000/month for OpenAI API access. And $800/month for Anthropic. Now you want another AI tool on top of that?” Team lead: ”…”
This is the double billing problem. Your company already pays for AI model access. Adding a per-seat AI platform on top means paying twice for the same underlying models. It’s the #1 budget objection to AI tool adoption.
BYOK (Bring Your Own Key) eliminates it entirely.
What BYOK actually means
BYOK lets you connect your existing API keys from any supported LLM provider directly into JieGou. When you run a recipe or workflow, JieGou sends the request to your provider using your key. You pay the provider directly at their published rates. JieGou adds zero markup on API costs.
Supported providers:
- Anthropic — Claude Opus, Sonnet, Haiku
- OpenAI — GPT-4o, GPT-4 Turbo, GPT-3.5
- Google — Gemini 2.5 Pro, Gemini 2.5 Flash
- Meta — Llama 3 (via Groq, Bedrock, or self-hosted)
- Mistral — Mistral Large, Medium, Small
- Cohere — Command R+, Command R
- Amazon Bedrock — all Bedrock-supported models
- Azure OpenAI — your Azure deployment endpoints
- Groq — ultra-fast inference for supported models
You can add multiple keys from multiple providers. Each recipe can use a different model. Your SEO analysis runs on Claude (strong at structured reasoning). Your creative copy runs on GPT-4o (strong at creative writing). Your quick summarization runs on Gemini Flash (fast and cheap).
How it works technically
When you add an API key to JieGou, it’s encrypted with AES-256-GCM before storage. The key never appears in logs, analytics, or error messages. When a recipe executes, JieGou decrypts the key in memory, sends the request to your provider, and discards the decrypted key immediately after the response.
Your data flows directly to the provider you chose. If you use your Anthropic key, the data goes to Anthropic’s API — subject to Anthropic’s data handling policies (which, for API usage, means they don’t train on your data). Same for OpenAI, Google, and every other provider.
JieGou never proxies your data through a third-party intermediary. There’s no “JieGou API” sitting between you and your provider. It’s a direct connection with your key.
The cost comparison
Let’s compare three scenarios for a 10-person marketing team running ~500 recipe executions per month.
Scenario 1: ChatGPT Teams
- Cost: $25/user/month = $250/month
- Models: GPT-4 only (with usage caps)
- Workflows: None (chat interface only)
- Analytics: Basic usage stats
- Total annual cost: $3,000
Scenario 2: AI platform with bundled models
- Cost: $40/user/month = $400/month (typical mid-market AI platform)
- Models: 2-3 models with platform markup on tokens
- Workflows: Basic
- Analytics: Some
- Total annual cost: $4,800
Scenario 3: JieGou + BYOK
- JieGou Pro: $29/month (flat, not per-user)
- API costs: ~500 executions x ~4,000 tokens avg x $3/million tokens = ~$6/month (Claude Haiku) to ~$40/month (Claude Opus)
- Models: 9 providers, unlimited model selection
- Workflows: Full engine with conditions, loops, approval gates, schedules
- Analytics: Complete execution tracking and ROI measurement
- Total annual cost: $420-$828
The BYOK approach costs 70-85% less than per-seat alternatives while providing more model choice, better workflows, and complete analytics.
Why BYOK is an adoption accelerator
Most people think of BYOK as a technical feature for developers. It’s actually the single most important adoption enabler for teams that already use AI.
It eliminates the budget conversation
When a team lead requests a new AI tool, finance asks “how much?” If the answer is “$25/user/month,” that’s a budget line item that needs approval. If the answer is “$29/month flat, and it uses the API keys we’re already paying for,” the budget conversation is over before it starts.
It proves you’re not paying twice
The double billing objection kills more AI tool purchases than any feature gap. BYOK makes it physically impossible to pay twice — you’re using the same keys, the same providers, the same billing relationship. JieGou just gives you a better interface for using them.
It lets you start free
JieGou’s free tier supports BYOK. A team can connect their existing API keys, install a department pack, and start running recipes without any new procurement process. By the time they want to upgrade to Pro for unlimited recipes, they already have ROI data to justify the $29/month.
It removes model lock-in fear
The #2 objection to AI tools (after cost) is “what if a better model comes out?” With BYOK, switching models takes 10 seconds. Add a new key, change the model selection on your recipes, done. No vendor negotiations. No contract changes. No data migration.
Setting up BYOK
The setup takes under a minute:
- Go to Settings → API Keys in your JieGou account
- Click Add Key and select your provider (e.g., Anthropic)
- Paste your API key
- JieGou validates the key with a test call and encrypts it
That’s it. Every recipe in your account can now use that provider’s models. You can set a default model for all recipes or override per-recipe.
To add a second provider, repeat the process. Most teams add 2-3 providers: one for high-quality output (Claude Opus or GPT-4), one for fast/cheap tasks (Gemini Flash or Claude Haiku), and one for comparison via bakeoffs.
BYOK + Bakeoffs: measure before you commit
Here’s a powerful pattern: use BYOK to add keys from multiple providers, then run JieGou’s bakeoff system to compare models side by side on your actual recipes.
A bakeoff runs the same recipe with the same inputs across 2-5 models simultaneously. You see the outputs side by side and score them. After a few rounds, you know exactly which model performs best for each recipe — based on your data, your prompts, your quality criteria.
This eliminates the “which model should we use?” question. You’re not reading benchmarks or blog posts. You’re measuring performance on your actual workloads.
The bottom line
BYOK reframes the AI platform purchase from “a new cost” to “a better way to use what you already pay for.” It turns a budget conversation into a no-brainer. It eliminates lock-in anxiety. And it lets teams start immediately without procurement.
If your company already has API keys from Anthropic, OpenAI, or Google, you can be running structured AI workflows inside JieGou within five minutes — at zero additional model cost.